ruffrecords
Well-known member
New version lunch box PSU boards arrived this morning. I am going to build one to make sure it is OK and fit it with a big external heatsink on the 12V before I put them in the white market.
Cheers
Ian
Cheers
Ian
Feeler: EZ Tube Lunch Box
- Thread starter ruffrecords
- Start date
Help Support GroupDIY Audio Forum:
ruffrecords
Well-known member
dipfrik said:Big changes? I'am in for 2!
Cheers, Sven
Just a few small ones. The main change was including some larger pins to wire to the LM317 regulator so it could be wired to one fitted to an off-board heatsink to cater for larger heater currents.
Cheers
Ian
ruffrecords
Well-known member
Now my op is over I want to get this project moving again. I have the new PSU PCBs so I'll be getting that sorted out. In the meantime I want to canvass opinion about input and output connectors. To keep costs down, it makes sense to use an off the shelf 19inch sub-rack as the main frame. These come in various depths. If you want to have all the connections at the rear and use all six slots for active circuits then you need a deep one, especially if you also will need to use the space for output transformers.. deep ones are more expensive and you will need to make a rear panel to hold the connectors etc.
An alternative is to use Holgers approach where the connections are at the front.. So you have say, four modules of active circuits and reserve two slots for I/O. You could also include output transformer here and possibly meters too. This means the sub-rack is less deep and hence cheaper and we only need one type of simple rear panel with either dc connector or even just a simple hole for a flying dc supply lead.
So, the question is, which do you prefer? I am inclined to follow Holger's lead as it simplifies things . We could then develop a range of alternative front panels for the two spare slots which people could get made locally or customise to their own needs.
Cheers
Ian
An alternative is to use Holgers approach where the connections are at the front.. So you have say, four modules of active circuits and reserve two slots for I/O. You could also include output transformer here and possibly meters too. This means the sub-rack is less deep and hence cheaper and we only need one type of simple rear panel with either dc connector or even just a simple hole for a flying dc supply lead.
So, the question is, which do you prefer? I am inclined to follow Holger's lead as it simplifies things . We could then develop a range of alternative front panels for the two spare slots which people could get made locally or customise to their own needs.
Cheers
Ian
ruffrecords
Well-known member
I am starting to get stuck into this. Ordered a couple of 3U racks today. I have set up a new blog to record progress:
http://eztubelunchbox.blogspot.co.uk/
Cheers
Ian
http://eztubelunchbox.blogspot.co.uk/
Cheers
Ian
ruffrecords
Well-known member
I have done some more thinking about the enclosure and added a third variant to the mechanical alternatives.
The full width 19 inch sub-rack could hold up to 6 modules and be used to make some serious tube designs. Even if you were to use a couple of slots for inputs/outputs at the front, you can still fit four modules. However there is no low cost alternative. So I have added a 3 module wide sub-rack to my order because I think this will satisfy the need for a lot of people to have just a couple of mic pres or one channel strip with EQ in a box. Obviously you cannot just stick a 3 module wide sub-rack into a 19 inch rack so I am assuming it will be a stand alone device, possibly finished with wooden sides. This is what I have added to the order.
I have also been thinking about the cost of various options. To put this into context, one thing we can never forget is tht there are high voltages in tube circuits and any lunch box needs to do its best to keep your fingers away from them for obvious reasons. Tube circuits are also generally high impedance so they can be susceptible to external interference. The solution to both these issues is screening.
The EZTube Mixer PCBs are not screened so the enclosure needs to provide screening. However, a regular sub-rack does not come with any screening. Most of the sub-racks I use I buy from a UK supplier called SRS. They have two types of sub-rack. Their low cost product, the Eurorack, uses a satin finish aluminium whic is not conductive. Top and bottom panels in the same material are available so these will protect you from high voltages but not provide very good screening. This is probably the lowest cost solution.
Their other sub-rack type is the Versi Rack. This is designed to provide EMC protection and uses aluminium that is made deliberately conductive. Its top and bottom screens are similarly conductive so this also provides both good screening and keeps you away from high voltages. This is the mid price solution.
The third and final alternative is to screen individual PCBs. This is Holger's method. Each 3U PCB is contained inside an aluminium case which plugs into the sub-rack. This is the most expensive option but provides the best screening of all.
I have not yet looked at sub-racks by other manufacturers and of course you can alsways buy a bare sub-rack and add your own screening.
Cheers
Ian
The full width 19 inch sub-rack could hold up to 6 modules and be used to make some serious tube designs. Even if you were to use a couple of slots for inputs/outputs at the front, you can still fit four modules. However there is no low cost alternative. So I have added a 3 module wide sub-rack to my order because I think this will satisfy the need for a lot of people to have just a couple of mic pres or one channel strip with EQ in a box. Obviously you cannot just stick a 3 module wide sub-rack into a 19 inch rack so I am assuming it will be a stand alone device, possibly finished with wooden sides. This is what I have added to the order.
I have also been thinking about the cost of various options. To put this into context, one thing we can never forget is tht there are high voltages in tube circuits and any lunch box needs to do its best to keep your fingers away from them for obvious reasons. Tube circuits are also generally high impedance so they can be susceptible to external interference. The solution to both these issues is screening.
The EZTube Mixer PCBs are not screened so the enclosure needs to provide screening. However, a regular sub-rack does not come with any screening. Most of the sub-racks I use I buy from a UK supplier called SRS. They have two types of sub-rack. Their low cost product, the Eurorack, uses a satin finish aluminium whic is not conductive. Top and bottom panels in the same material are available so these will protect you from high voltages but not provide very good screening. This is probably the lowest cost solution.
Their other sub-rack type is the Versi Rack. This is designed to provide EMC protection and uses aluminium that is made deliberately conductive. Its top and bottom screens are similarly conductive so this also provides both good screening and keeps you away from high voltages. This is the mid price solution.
The third and final alternative is to screen individual PCBs. This is Holger's method. Each 3U PCB is contained inside an aluminium case which plugs into the sub-rack. This is the most expensive option but provides the best screening of all.
I have not yet looked at sub-racks by other manufacturers and of course you can alsways buy a bare sub-rack and add your own screening.
Cheers
Ian
I was planning on using this screening panel from schroff, http://canada.newark.com/schroff/24560-081/emc-cvr-bolt-on-84hp-x-355/dp/64T7511 Not cheap but probably cheaper then the individual cassettes.
Attachments
Ok, maybe it is too late, but after a few days in front, behind and under 
Holger's console there are some thoughts in my mind i want too share. Basically maybe we should start thinking of a much bigger depth of modules. And as far i can grab from the Fischer catalog, we can go 235,5mm maybe 295,5mm deep.
Which means a lot more space for a lot of things i would put into a module.
1. self contained modules -> everything is inside a module. Which means at least including the output transformer. There is definitely space for an output transformer and call me crazy, i think you can put three inside to support 2 additional transformers (send & return) for an insert path.
2. no external psu ( API can do it, perhaps we too ) -> i wont give up the dream of a 2 module wide (28HP) integrated PSU. Is it possible? I feel very strong! I have tested this in Target (3D mode) the last few hours and at the end i rebuild the complete Lunchbox PSU from Ian including a toroidal transformer inside the module on a 220mmx100mm pcb space. By the way with an additional 12VDC-Utility support for relays, led and more. I know we have the regulator cooling topic, but Fischer has also modules with dedicated cooling fins outside for mounting.
I have attached a few screens from my 1st exploration. The "dummy" toroidal transformer has a 90mm diameter by a 50mm height so far, which actually match a few of the bigger one here in my lab. If its not fit horizontal, we can mount the transformer vertical and build the pcb around.
3. the "patchbay" IO-module -> these are the moments i love sitting right next to Holger the last weeks! Let us think for a moment: if we have totally self contained modules (which theoretically also means that an EQ has at least one tube gain stage for compensating gain loss and input/output transformers), we should have a way to freely assign the audio signal path (DSub 25 connector preferred at the backside for i/o like in/out, inserts and additional 2-4 XLR jacks at the front for Mic's) between the available modules. I really love the freedom of choice. If i need an eq after the preamp - patch it. Or only want to use the "tube eq" and the preamp standalone on a specific input signal - patch it. Maybe here is the better place for possible send & return functionality with transformers, buffer amps etc. i mentioned under 1). The "fader"? I think it should be there! (I remember someone mentioned "Radial Workhorse" racks here previously or?)
I know, only 3 modules left to fill in a 19" rack. But i love the idea! And if we have a Tube compessor in the future we have a perfect chain -> preamp -> eq <-> comp.
At the end with deeper modules you can also think of putting two preamps into one 3U module. Then you can think of the following chain:
4 drum mics "patched" into 2x2 preamp modules (1) and (2) -> 3rd module is a twinline amp module -> 4th module is the "patchbay" IO-module which give me access to all the four preamps for attentuation/mixing. You need an external stereo bus compressor - patch it! - it is abvailable over the DSub 25 connector at the backside.
So maybe it is really too late, or i'am only overstate the added value of deeper modules.
But it is fun to think about it!
What do you think?
Night, Sven
Holger's console there are some thoughts in my mind i want too share. Basically maybe we should start thinking of a much bigger depth of modules. And as far i can grab from the Fischer catalog, we can go 235,5mm maybe 295,5mm deep.Which means a lot more space for a lot of things i would put into a module.
1. self contained modules -> everything is inside a module. Which means at least including the output transformer. There is definitely space for an output transformer and call me crazy, i think you can put three inside to support 2 additional transformers (send & return) for an insert path.
2. no external psu ( API can do it, perhaps we too
) -> i wont give up the dream of a 2 module wide (28HP) integrated PSU. Is it possible? I feel very strong! I have tested this in Target (3D mode) the last few hours and at the end i rebuild the complete Lunchbox PSU from Ian including a toroidal transformer inside the module on a 220mmx100mm pcb space. By the way with an additional 12VDC-Utility support for relays, led and more. I know we have the regulator cooling topic, but Fischer has also modules with dedicated cooling fins outside for mounting.I have attached a few screens from my 1st exploration. The "dummy" toroidal transformer has a 90mm diameter by a 50mm height so far, which actually match a few of the bigger one here in my lab. If its not fit horizontal, we can mount the transformer vertical and build the pcb around.
3. the "patchbay" IO-module -> these are the moments i love sitting right next to Holger the last weeks! Let us think for a moment: if we have totally self contained modules (which theoretically also means that an EQ has at least one tube gain stage for compensating gain loss and input/output transformers), we should have a way to freely assign the audio signal path (DSub 25 connector preferred at the backside for i/o like in/out, inserts and additional 2-4 XLR jacks at the front for Mic's) between the available modules. I really love the freedom of choice. If i need an eq after the preamp - patch it. Or only want to use the "tube eq" and the preamp standalone on a specific input signal - patch it. Maybe here is the better place for possible send & return functionality with transformers, buffer amps etc. i mentioned under 1). The "fader"? I think it should be there! (I remember someone mentioned "Radial Workhorse" racks here previously or?)
I know, only 3 modules left to fill in a 19" rack. But i love the idea! And if we have a Tube compessor in the future
we have a perfect chain -> preamp -> eq <-> comp.At the end with deeper modules you can also think of putting two preamps into one 3U module. Then you can think of the following chain:
4 drum mics "patched" into 2x2 preamp modules (1) and (2) -> 3rd module is a twinline amp module -> 4th module is the "patchbay" IO-module which give me access to all the four preamps for attentuation/mixing. You need an external stereo bus compressor - patch it! - it is abvailable over the DSub 25 connector at the backside.
So maybe it is really too late, or i'am only overstate the added value of deeper modules.
But it is fun to think about it!
What do you think?
Night, Sven
ruffrecords
Well-known member
Hi Sven,
Very interesting post, particularly because you have a first and appreciation of the actually building a mixer. Taking your point one by one.
1. Self contained modules. I agree bigger modules are a good idea. The question is which way to make them bigger? You can go deeper as you suggest but you do not get any more front panel space and there is very little room in a 3U front panel, especially if you want to fit in mic pre and EQ. That's why in the Mark 3 I am making 6U high PCBs of the same depth as the EZTube Mixer boards. You get more than twice the PCB area and there is plenty of room to fit an output transformer for example. The other problem is that the EZTubeMixer 32 way socket has no spare pins for extra connections like balanced outs and insert points, but a 6U PCB has room for an extra 32 way connector if you need it. The very first mixer I sold used 6U high PCBs and this meant there was enough room to fit a local regulator for the heaters. This saves having a big heatsink in the power supply. I tried to include this when I designed the EZTube boards but there was not enough room. A deeper board would allow this to be included. The standard depths start at 100mm and increase in 60mm increments so the next is 160mm, which is what we use, then 220mm and then 280mm. I guess a 200mm depth might just get enough room for an output transformer and a heater regulator. Mmmmm..... a 6U board solves all the problems with ease but it is 6U high. A 22mm deep 3U card might just solve the problems we need to solve. Except I am not sure we have a problem. At the moment the cards have an input transformer and no output transformer. The idea of using the deeper rack option is that there is enough room at the back to mount the output transformers if you want them. If you don't need them then you can use the shorter rack. with a deeper card you always need the deeper rack even if you do not fit the output transformers.
2. I agree there is something quite attractive about an internal power supply. As you know I tried some time ago to squeeze it all onto one card without success. However, it occurs to me that ,for the lunch box, the phantom supply is over specified - it can supply up to 100mA at 48V. If we were to reduce that to 20mA for the lunch box then it should be possible to reduce the heatsink, hopefully enough to fit it onto a standard Eurocard. I will have a look at that. I am still unsure about having the transformer inside the lunch box as I had quite a bit of trouble with this in the EZTubeMixer prototype which has built in power supplies. If the power supply PCB could be reduced to a Eurocard size then at least we have still the option to fit the transformer inside or outside. Of course, if we can get heater regulation onto the active PCBs then this removes the heater heatsink altogether and makes things much easier.
3. The great thing about Holger is that, not only is he a builder, he is a user. He has made several additions of his own the the EZTube project - he came up with the 2 module motherboard with integrated relays idea and the front panel I/O. I agree that a DB25 at the rear could be very handy. That can be included with the present system at the rear along with any necessary output transformers. The question of making the lunch box flexible is a big one. I think the main missing link is a self contained EQ module with its own input transformer and gain make up amplifier and (possibly) output transformer. Trouble is the EQ parts take up a lot of room. I did design a buffer amplifier stage - Holger has some of these - but it needs a 6V eater supply in addition to the 12V one we already have which complicates matters, The other problem is it does not have enough gain to act as gain make up for some EQs and still maintain low distortion figures. Once again, the 6U format solves this easily. There is plenty of room for the EQ and the amplifier and an output transformer.
In conclusion, the 6U PCB is probably the best long term solution but it does not make a nice convenient size for a lunch box. The power supply is definitely worth looking at again especially if we could get 12V regs onto the active boards. Now, I know I said there is no room but.... I h ve been experimenting with replacing the large 4,7uF film output capacitors with much smaller footprint 47uF electrolytics. I know electrolytics in the signal path are heresy to some people but in the test I have done so far there is no measurable degradation in performance and the low frequency response is significantly improved. So, replacing the output caps with electrolytics might free ip enough space for a 12V on board reg which in turn simplifies the power supply which means it might fit onto a standard Eurocard.
Definitely something to think about.
Cheers
Ian
Very interesting post, particularly because you have a first and appreciation of the actually building a mixer. Taking your point one by one.
1. Self contained modules. I agree bigger modules are a good idea. The question is which way to make them bigger? You can go deeper as you suggest but you do not get any more front panel space and there is very little room in a 3U front panel, especially if you want to fit in mic pre and EQ. That's why in the Mark 3 I am making 6U high PCBs of the same depth as the EZTube Mixer boards. You get more than twice the PCB area and there is plenty of room to fit an output transformer for example. The other problem is that the EZTubeMixer 32 way socket has no spare pins for extra connections like balanced outs and insert points, but a 6U PCB has room for an extra 32 way connector if you need it. The very first mixer I sold used 6U high PCBs and this meant there was enough room to fit a local regulator for the heaters. This saves having a big heatsink in the power supply. I tried to include this when I designed the EZTube boards but there was not enough room. A deeper board would allow this to be included. The standard depths start at 100mm and increase in 60mm increments so the next is 160mm, which is what we use, then 220mm and then 280mm. I guess a 200mm depth might just get enough room for an output transformer and a heater regulator. Mmmmm..... a 6U board solves all the problems with ease but it is 6U high. A 22mm deep 3U card might just solve the problems we need to solve. Except I am not sure we have a problem. At the moment the cards have an input transformer and no output transformer. The idea of using the deeper rack option is that there is enough room at the back to mount the output transformers if you want them. If you don't need them then you can use the shorter rack. with a deeper card you always need the deeper rack even if you do not fit the output transformers.
2. I agree there is something quite attractive about an internal power supply. As you know I tried some time ago to squeeze it all onto one card without success. However, it occurs to me that ,for the lunch box, the phantom supply is over specified - it can supply up to 100mA at 48V. If we were to reduce that to 20mA for the lunch box then it should be possible to reduce the heatsink, hopefully enough to fit it onto a standard Eurocard. I will have a look at that. I am still unsure about having the transformer inside the lunch box as I had quite a bit of trouble with this in the EZTubeMixer prototype which has built in power supplies. If the power supply PCB could be reduced to a Eurocard size then at least we have still the option to fit the transformer inside or outside. Of course, if we can get heater regulation onto the active PCBs then this removes the heater heatsink altogether and makes things much easier.
3. The great thing about Holger is that, not only is he a builder, he is a user. He has made several additions of his own the the EZTube project - he came up with the 2 module motherboard with integrated relays idea and the front panel I/O. I agree that a DB25 at the rear could be very handy. That can be included with the present system at the rear along with any necessary output transformers. The question of making the lunch box flexible is a big one. I think the main missing link is a self contained EQ module with its own input transformer and gain make up amplifier and (possibly) output transformer. Trouble is the EQ parts take up a lot of room. I did design a buffer amplifier stage - Holger has some of these - but it needs a 6V eater supply in addition to the 12V one we already have which complicates matters, The other problem is it does not have enough gain to act as gain make up for some EQs and still maintain low distortion figures. Once again, the 6U format solves this easily. There is plenty of room for the EQ and the amplifier and an output transformer.
In conclusion, the 6U PCB is probably the best long term solution but it does not make a nice convenient size for a lunch box. The power supply is definitely worth looking at again especially if we could get 12V regs onto the active boards. Now, I know I said there is no room but.... I h ve been experimenting with replacing the large 4,7uF film output capacitors with much smaller footprint 47uF electrolytics. I know electrolytics in the signal path are heresy to some people but in the test I have done so far there is no measurable degradation in performance and the low frequency response is significantly improved. So, replacing the output caps with electrolytics might free ip enough space for a 12V on board reg which in turn simplifies the power supply which means it might fit onto a standard Eurocard.
Definitely something to think about.
Cheers
Ian
Hey Ian,
thanks for answering my thoughts in such a detail!
You are absolutely right, that building or better thinking in mixing environments definitely drives me a lot in the past few weeks after getting deeper and deeper in all the various options that are possible or at least needed.
I challenged myself a lot with different setups and architectures and from my actual perspective there are two ways and a lot of "open fields" between them.
A) The EZ-Tube standalone Projects -> I call them "a tube channel strip in a box". Everything is available as PCB's at the Emporium and well documented. You can combine a preamp with an EQ of your choice, put it into a case of your imagination and be (very)happy. And that's what i did over the past few month and will be continue to follow this on a side project.
And then there is B)...
B) ...the idea of thinking in "repeating channels" in larger environments. At this point you need to think so different, that sometimes it drives me crazy to fullfill at least the minimal requirements of a mixer. Holger's actual mixing console shows me that in very positive and incentive way.
Between A) and B) i think there is nothing that makes sense for me right now. And to be clear, there is no need to think in numbers like 4x2 or bigger ("full blown tube mixer"). Two channels at the latest and you are asking yourself exactly the right questions. What is my real world usage of two or later more ez-tube channels pimped with some eq's of your choice, fancy vu meters, faders or knobs for attentuation, you name it?
At this point for me the EZ Tube Lunchbox Project should be more than a "testing rig" for "modules", it should give everyone from one to x channels/modules a solid foundation they can learn, build and grow on.
Sorry, for the long intro Back to technical stuff!
"Solving problems and requirements with bigger modules"
I couldn't agree more that going deeper is only half of the solutions if we keep in mind, that nothing changed in terms of real estate on the front panel. So why not starting finally and ultimately to think in 6U and 220mm maybe 280mm depth? It makes things so much easier on the road. And you are right, the Mark 3 is exactly the logical next step out of all your experience over the last years and will be the best long term solution i see so far.
Going deeper, i would like to define some constraints in terms of PCB & Panel design and requirements, which everyone (maybe myself 1st) can rely on. For instance: Two 32 way connectors, with a defined pinout to start with. A PSU Module everyone can slide in and focus on the much more interesting things.
It should be like designing a 500 module. I know my constraints in terms of pinout, size of the pcb, panel with an height and so on. What at the end comes out of any module is defined by the creator/builder himself beside the question of a real world usage.
Last but not least, i like the idea of distributing the load of work accoss several shoulders. With a set of "rules", i think we can grow much faster the idea of a lunchbox and quicker solving problems that will be arise on this journey.
So what are the logical next steps to bring this into a 1st real world "Box" that is standing on my workbench and i can play with?
Cheers, Sven
PS: Is putting the regulators onto the active boards a common approach?
thanks for answering my thoughts in such a detail!
You are absolutely right, that building or better thinking in mixing environments definitely drives me a lot in the past few weeks after getting deeper and deeper in all the various options that are possible or at least needed.
I challenged myself a lot with different setups and architectures and from my actual perspective there are two ways and a lot of "open fields" between them.
A) The EZ-Tube standalone Projects -> I call them "a tube channel strip in a box". Everything is available as PCB's at the Emporium and well documented. You can combine a preamp with an EQ of your choice, put it into a case of your imagination and be (very)happy. And that's what i did over the past few month and will be continue to follow this on a side project.
And then there is B)...
B) ...the idea of thinking in "repeating channels" in larger environments. At this point you need to think so different, that sometimes it drives me crazy to fullfill at least the minimal requirements of a mixer.
Holger's actual mixing console shows me that in very positive and incentive way.Between A) and B) i think there is nothing that makes sense for me right now. And to be clear, there is no need to think in numbers like 4x2 or bigger ("full blown tube mixer"). Two channels at the latest and you are asking yourself exactly the right questions. What is my real world usage of two or later more ez-tube channels pimped with some eq's of your choice, fancy vu meters, faders or knobs for attentuation, you name it?
At this point for me the EZ Tube Lunchbox Project should be more than a "testing rig" for "modules", it should give everyone from one to x channels/modules a solid foundation they can learn, build and grow on.
Sorry, for the long intro
Back to technical stuff!"Solving problems and requirements with bigger modules"
I couldn't agree more that going deeper is only half of the solutions if we keep in mind, that nothing changed in terms of real estate on the front panel. So why not starting finally and ultimately to think in 6U and 220mm maybe 280mm depth? It makes things so much easier on the road. And you are right, the Mark 3 is exactly the logical next step out of all your experience over the last years and will be the best long term solution i see so far.
Going deeper, i would like to define some constraints in terms of PCB & Panel design and requirements, which everyone (maybe myself 1st
) can rely on. For instance: Two 32 way connectors, with a defined pinout to start with. A PSU Module everyone can slide in and focus on the much more interesting things. It should be like designing a 500 module. I know my constraints in terms of pinout, size of the pcb, panel with an height and so on. What at the end comes out of any module is defined by the creator/builder himself beside the question of a real world usage.
Last but not least, i like the idea of distributing the load of work accoss several shoulders. With a set of "rules", i think we can grow much faster the idea of a lunchbox and quicker solving problems that will be arise on this journey.
So what are the logical next steps to bring this into a 1st real world "Box" that is standing on my workbench and i can play with?
Cheers, Sven
PS: Is putting the regulators onto the active boards a common approach?
ruffrecords
Well-known member
Hi Sven,
Thanks for your interesting reply. I will read it carefully and reply to it. In the meantime, one of your wishes is about to become true. I spent much of today trying to squeeze the power supply into a 3U PCB and I think I succeeded. Here is a pic of the layout I came up with:
You need to make some hand wired connections to the 32 way connector for the 12.6V and 48V supplies but that should be no problem. To get the supply into the space I have:
1. Squeezed the components in the 12.6V supply to make space for the HT supply below.
2. Moved a couple of resistors in the HT supply to avoid the 12.6V heatsink.
3. Deleted all the input and output terminal blocks and replaced them with chunky pads to hard wire to.
4. Reduced the size of the 48V smoothing cap from 35mm diameter to 25mm diameter. This means the biggest 100V value we are likely to get is 1000uF instead of the 4700uF we use now. This means the ripple will be 5 times higher for a given current draw than with the original design. This may or may not be an issue but it should certainly be as good up to currents of 20mA and will probably be OK for twice that amount.
5. Deleted the heatsink for the 48V supply. If we take less current we generate less heat. I have not actually specified a heatsink but there is plenty of room to fit a large flat piece of aluminium for example. You just need to be aware it is at 48V so you must not let it touch the chassis.
The PCB needs some tidying up and I need to change the component idents to match the original power supply PCB so you can use the same schematic and BOM for both. Then I will get some prototype boards made.
I am assuming this fits into a double width module with the transformer occupying the other half and connected to the PCB by flying leads. I am not sure if Fischer do a 28HP 3U module but if they do it would be ideal. The only thing I have not thought about is where the mains gets in.I guess you could mount some kind of connector on the rear of the module behind the transformers for this. Alternatively you could mount the transformers on the side of the sub rack itself and wire to the PCB. The only problem I see with that is you cannot remove the PCB. That would mean going back to plug/sockets which take up space.
Let me know what you think.
Cheers
Ian
Cheers
Ian
P.S I just checked at Buerklin and they have a 28HP module.
Thanks for your interesting reply. I will read it carefully and reply to it. In the meantime, one of your wishes is about to become true. I spent much of today trying to squeeze the power supply into a 3U PCB and I think I succeeded. Here is a pic of the layout I came up with:
You need to make some hand wired connections to the 32 way connector for the 12.6V and 48V supplies but that should be no problem. To get the supply into the space I have:
1. Squeezed the components in the 12.6V supply to make space for the HT supply below.
2. Moved a couple of resistors in the HT supply to avoid the 12.6V heatsink.
3. Deleted all the input and output terminal blocks and replaced them with chunky pads to hard wire to.
4. Reduced the size of the 48V smoothing cap from 35mm diameter to 25mm diameter. This means the biggest 100V value we are likely to get is 1000uF instead of the 4700uF we use now. This means the ripple will be 5 times higher for a given current draw than with the original design. This may or may not be an issue but it should certainly be as good up to currents of 20mA and will probably be OK for twice that amount.
5. Deleted the heatsink for the 48V supply. If we take less current we generate less heat. I have not actually specified a heatsink but there is plenty of room to fit a large flat piece of aluminium for example. You just need to be aware it is at 48V so you must not let it touch the chassis.
The PCB needs some tidying up and I need to change the component idents to match the original power supply PCB so you can use the same schematic and BOM for both. Then I will get some prototype boards made.
I am assuming this fits into a double width module with the transformer occupying the other half and connected to the PCB by flying leads. I am not sure if Fischer do a 28HP 3U module but if they do it would be ideal. The only thing I have not thought about is where the mains gets in.I guess you could mount some kind of connector on the rear of the module behind the transformers for this. Alternatively you could mount the transformers on the side of the sub rack itself and wire to the PCB. The only problem I see with that is you cannot remove the PCB. That would mean going back to plug/sockets which take up space.
Let me know what you think.
Cheers
Ian
Cheers
Ian
P.S I just checked at Buerklin and they have a 28HP module.
Hey Ian,
i see this 3U psu module coming! I will check your layout later in detail.
I also quickly double checked the fischer catalog and 3U/28HP should be available.
Because i only found the definition for a max. 100x220mm eurocard, means we should think in a this max. depth (228mm) of a module?
Update: uppps...you squeezed it onto a 100x160 eurocard?
Update: PDF from www.buerklin.com: https://www.buerklin.com/datenblaetter/H872250_TD.pdf?ch=48237
The transformer should be sitting inside the module, but a very good point you mentioned was the input for the mains. I totally did not take that into account.
It shouldn't be a problem to reserve space for an IEC connector at the back panel of the module, but did we not have a problem with the backplane pcb in terms of providing space for the power cord the reach the IEC connector?
Cheers, Sven
i see this 3U psu module coming!
I will check your layout later in detail.I also quickly double checked the fischer catalog and 3U/28HP should be available.
Because i only found the definition for a max. 100x220mm eurocard, means we should think in a this max. depth (228mm) of a module?
Update: uppps...you squeezed it onto a 100x160 eurocard?
Update: PDF from www.buerklin.com: https://www.buerklin.com/datenblaetter/H872250_TD.pdf?ch=48237
The transformer should be sitting inside the module, but a very good point you mentioned was the input for the mains. I totally did not take that into account.
It shouldn't be a problem to reserve space for an IEC connector at the back panel of the module, but did we not have a problem with the backplane pcb in terms of providing space for the power cord the reach the IEC connector?
Cheers, Sven
ruffrecords
Well-known member
dipfrik said:Update: uppps...you squeezed it onto a 100x160 eurocard?
YES!
The transformer should be sitting inside the module, but a very good point you mentioned was the input for the mains. I totally did not take that into account.
It shouldn't be a problem to reserve space for an IEC connector at the back panel of the module, but did we not have a problem with the backplane pcb in terms of providing space for the power cord the reach the IEC connector?
Cheers, Sven
That is a good point. It means we need a motherboard for an odd number of modules. So for a full width 19 inch sub-rack it would need to be 5 modules wide and for the half width version it needs to be 3 modules wide. As we can make 5 from 3 + 2 and we already have the 2 we just need a 3 module motherboard. I will look into that.
Cheers
Ian
Maybe we can make a modification to the motherboard and provide a large cutout, which than gives us the space for example the power cord to the PSU module backpanel? ...besides the effect of eventually more air circulation to and from the modules?
Having different motherboard pcb's makes it always more difficult to maintain over future revisions?
Something like this?
In general, are there any new ideas or feature requests for the motherboard PCB's based on the experience from you and Holger? I know Holger has developed his own 2 module backplane. Is thinking in 6U modules with two 32 way sockets a game changer here?
Holy moly...a lot of questions
Cheers, Sven
ruffrecords
Well-known member
dipfrik said:
I think there are tracks on the other side of the motherboard where you have drawn the holes. But I think we have another problem anyway; the PSU PCB plugs in. If it is in a module (which we want) then the whole module must plug in including the mains. If the mains is on a flying lead that you connect up after you plug in the module, then there is not much point in the PSU/module being plug-in - we need a mains plug in connector which is dangerous territory. We might just as well hard wire in the whole module including the dc connections and delete the 32 way connector on the PSU board. In other words have a fixed power supply a bit like the API lunch box.
In general, are there any new ideas or feature requests for the motherboard PCB's based on the experience from you and Holger? I know Holger has developed his own 2 module backplane. Is thinking in 6U modules with two 32 way sockets a game changer here?
Holy moly...a lot of questions
Cheers, Sven
Both Holger and I abandoned the 6 module motherboard and went for a 2 module version. Now, after wiring up a mixer with the 2 module version, I am thinking about using the 6 again as it simplifies wiring a lot. There's an error in the current 6 module one that needs fixing). Maybe we just do a special 4 module motherboard for the lunch box. 4 is a handy number. Two of them do 8 channels ( the 8 tracker) and 3 of them do 12 channels (Mark 3) plus a 2 module one for the last two. Mmmmmm.......
I have not had any requests for special features on the motherboard. However, in the bigger mixers there may be a need for three separate motherboards. One is the current EZTube Mixer one. The second would be for the 6U extra connector - the pins on this are not yet assigned. The third one could be a for a separate routing module. Let me explain:
I want to make custom Mark 3 mixers as easy as possible to build. So I aim to have a set of standard channel modules that are just mic pres with EQ, balanced in and out. There would be no pan controls or aux sends like on the EZTube Demo Mixer. Instead, all routing controls would be in a separate 3U routing module. This would have its own dedicated motherboard consisting mainly of buses and some dc for relay switching and bulbs.
This is very much the way we used to do it at Neve. A few different channel amps and lots of different routing modules.
Cheers
Ian
I think we can live without plugging in the PSU module directly to the motherboard. Providing a proper IEC connector at the module and not messing around with flying power cord wires is always a good idea. Getting the rails onto the motherboard can be surely archieved with som e Harting stuff. There is also a nice side effect in a decoupled PSU module. Sliding the PSU out of the rack is an option so the PSU module can be used during developing or building other modules outside the rack. I can remember that Holger has some kind of "debugger" module which brings one motherboard connector to the outside world with a flying cable. But this is future stuff.
I couldn't find any design notes or something like that at the Emporium right?
We talk about a kind of "supporter motherboard" if we building 6U modules or?
This is exactly what i wanted to bring onto the table with the "patchbay module" idea!
I really like this not only because it's fun to develop in terms of technical options, schematics and pcb design
but also it makes totally sense to me that a standard channel (micpre+eq) most of the time shouldn't care so much about what's happening with signals after it leaves the module.The actual design of the 3UPSU doesn't provide any VDC-Utility rails?
What the idea of providing e.g. a 12VDC Utility support? I'am not an expert, but some kind of "tapping" into the HTR is not a good idea or?
But we should provide this option inside a module if needed? Or is this also the "job" of the routing module in the future? Having Pan & Auxes separated from the micpre feels strange right now.
Cheers, Sven
ruffrecords
Well-known member
Hi Sven,
Agreed about the PSU. We can just fit and IEC and a dc connector of some sort at the back. I have not used the Harting connectors but I know Holger has. We can decide details later. As you say, it makes a nice little stand alone power supply too. I guess we will need to find room on the front panel for an on/off switch and a fuse and an indicator lamp.
The PSU design does not include any utility rails. I think for the lunch box we can use the 12.6V heater supply. With the power supply on board there is room for four modules. Even if all four modules are have tubes, the maximum heater draw per module is 0.45 amps so a total of 1.8 amps is required. I am sure the supply could provide 2.2 amps which leaves 400mA for utilities. Is that enough?
The error on the 6U motherboard is a mechanical one. The connectors are about 0.2mm from where they should be. Most times it makes no difference but that is why I no longer stock that board. I made the same mistake when I designed the two module motherboard which is why it is now at V2 with the mistake fixed. I still like the idea of a 4 module version. It should be easy to design by concatenating a pair of 2 module boards together.
For the Mark 3 6U modules, there are several options:
1. Simple mixers like the 8 tracker do not have room for a separate routing module so the pan and AUX send controls will be in the channel module just like in the EZ Tube Mixer.
2. If the module has a built in output transformer then this could use the second connector. However, if there is nothing else needed on this connector then it it a bit of a waste so I am thinking about re-assigning a couple of pins on the top 32 way connector.
3. Some modules will not have an output transformer. One example would be a twin EQ module. If you create a 6U version of the twin line amp then, as there are no mic pre switches/buttons, there is room to fit two identical EQs side by side. This build on the Mark 3 idea of having a daughter board to carry EQ. If the EQ is simple enough, e.g REDD EQ, there is even enough room for a pan and a couple of AUX sends. This makes a very compact line level only dual channel module. 8 of these would form the core of a nice 16 in to 2 mix down console. This module does not need the second 32 way connector.
4. Big mixers have channel amps with mic pre and EQ only. Pans, AUXes and routing are all in a separate routing module.
Bottom line is I am trying to out together the components that will make all the above options possible. I need to add a Mark 3 directory to my web site with some pics.
Cheers
Ian
Agreed about the PSU. We can just fit and IEC and a dc connector of some sort at the back. I have not used the Harting connectors but I know Holger has. We can decide details later. As you say, it makes a nice little stand alone power supply too. I guess we will need to find room on the front panel for an on/off switch and a fuse and an indicator lamp.
The PSU design does not include any utility rails. I think for the lunch box we can use the 12.6V heater supply. With the power supply on board there is room for four modules. Even if all four modules are have tubes, the maximum heater draw per module is 0.45 amps so a total of 1.8 amps is required. I am sure the supply could provide 2.2 amps which leaves 400mA for utilities. Is that enough?
The error on the 6U motherboard is a mechanical one. The connectors are about 0.2mm from where they should be. Most times it makes no difference but that is why I no longer stock that board. I made the same mistake when I designed the two module motherboard which is why it is now at V2 with the mistake fixed. I still like the idea of a 4 module version. It should be easy to design by concatenating a pair of 2 module boards together.
For the Mark 3 6U modules, there are several options:
1. Simple mixers like the 8 tracker do not have room for a separate routing module so the pan and AUX send controls will be in the channel module just like in the EZ Tube Mixer.
2. If the module has a built in output transformer then this could use the second connector. However, if there is nothing else needed on this connector then it it a bit of a waste so I am thinking about re-assigning a couple of pins on the top 32 way connector.
3. Some modules will not have an output transformer. One example would be a twin EQ module. If you create a 6U version of the twin line amp then, as there are no mic pre switches/buttons, there is room to fit two identical EQs side by side. This build on the Mark 3 idea of having a daughter board to carry EQ. If the EQ is simple enough, e.g REDD EQ, there is even enough room for a pan and a couple of AUX sends. This makes a very compact line level only dual channel module. 8 of these would form the core of a nice 16 in to 2 mix down console. This module does not need the second 32 way connector.
4. Big mixers have channel amps with mic pre and EQ only. Pans, AUXes and routing are all in a separate routing module.
Bottom line is I am trying to out together the components that will make all the above options possible. I need to add a Mark 3 directory to my web site with some pics.
Cheers
Ian
Don't know at what stage this project is right now but I've been looking at this and the tube mixer thread from time to time 'cause I really like your projects.
In the beginning of this thread I posted some pics of my Edyne lunchbox and it strikes me again how small and compact it is.
It's my understanding that the tallest part of the modules are the tubes, and that determines the module width?
What if you made a sub board for the tube sockets so the tube can be mounted along with the PCB like the Edyne 201-A? I've seen this in some tube mic projects here.
I guess this will make the cost rise but maybe the more compact every module is, the more attractive it gets??
To make it more appealing maybe the sub boards could be incorperated in the EZ Tube Mixer boards as well in the future. But then you would have to redo the layout of all those boards....hmmm....
Well, it was just a thought. Keep up the good work!!
Best regards
//Magnus
In the beginning of this thread I posted some pics of my Edyne lunchbox and it strikes me again how small and compact it is.
It's my understanding that the tallest part of the modules are the tubes, and that determines the module width?
What if you made a sub board for the tube sockets so the tube can be mounted along with the PCB like the Edyne 201-A? I've seen this in some tube mic projects here.
I guess this will make the cost rise but maybe the more compact every module is, the more attractive it gets??
To make it more appealing maybe the sub boards could be incorperated in the EZ Tube Mixer boards as well in the future. But then you would have to redo the layout of all those boards....hmmm....
Well, it was just a thought. Keep up the good work!!
Best regards
//Magnus
ruffrecords
Well-known member
Dr Gris said:
My very early designs used the 6CG7 tube which is taller than the ones used in the EZ Tube Mixer. I mounted them width ways across the modules. This made them 4.2 inches wide so I had to build two channels in each module else they would have been impractically wide.
When I started thinking about the EZ Tube Mixer I was still using the 6CG7 tube and my first thought was to mount them exactly as you suggest. The trouble is the they occupied so much of the available volume that there was hardly any room for other components. I thought of mounting components underneath the tubes but they would have got too hot, In the end I decided mounting tubes across the modules was the best option but the way forward was to look for a design using shorter tubes. The result is the EZ Tube Mixer.
Cheers
Ian
Hi Ian,
let me 1st going back to the 3U-PSU mechanicals or better checking the clearance between a toroidal transformer and the pcb. I started with a 3D Model which represents 1:1 the available volume including the eurocard 100x160mm and the transformer.
Here are some shots. My 1st impression was, that the transformer (90mm diameter by 50mm height) will determines the actual layout of the pcb. The transfomer dimensions are actual my personal assumption, so we should check if our requirements can be archieved within the limits.
Which brings me shortly back to the 12VDC-Utility rail. After a few minutes (ok it was way longer) of discussion with Holger today, i definitely believe that we need more than 400mA and should not interfere with HT or HTR rails. Which brings me to the current secondaries:
I: HT -> 240VDC (300mA)
II HTR -> 14VDC (2A)
II: 48V -> 50VAC (20mA) <-?
IV: Utility -> 14DVC (>= 1A)
Again i'am not an expert to precise dimensioning the final specs, but can we really get 4 secondaries with such specs in a 90x50mm block?
I also took some real world photos with a different orientation of the transformer. Interestingly in this scenario there is a lot more space. So much space, that i even could fit a second transformer inside.
Bottom line: Right now i think we should 1st find the right transformer and then derives the final pcb design for the psu from there?
What do you think?
Cheers, Sven
let me 1st going back to the 3U-PSU mechanicals or better checking the clearance between a toroidal transformer and the pcb. I started with a 3D Model which represents 1:1 the available volume including the eurocard 100x160mm and the transformer.
Here are some shots. My 1st impression was, that the transformer (90mm diameter by 50mm height) will determines the actual layout of the pcb. The transfomer dimensions are actual my personal assumption, so we should check if our requirements can be archieved within the limits.
Which brings me shortly back to the 12VDC-Utility rail. After a few minutes (ok it was way longer
) of discussion with Holger today, i definitely believe that we need more than 400mA and should not interfere with HT or HTR rails. Which brings me to the current secondaries:I: HT -> 240VDC (300mA)
II HTR -> 14VDC (2A)
II: 48V -> 50VAC (20mA) <-?
IV: Utility -> 14DVC (>= 1A)
Again i'am not an expert to precise dimensioning the final specs, but can we really get 4 secondaries with such specs in a 90x50mm block?
I also took some real world photos with a different orientation of the transformer. Interestingly in this scenario there is a lot more space. So much space, that i even could fit a second transformer inside.
Bottom line: Right now i think we should 1st find the right transformer and then derives the final pcb design for the psu from there?
What do you think?
Cheers, Sven
