mattamatta
Well-known member
JohnRoberts said:
I cut open a small radial cap once - something like a ceramic type - and was surprised to see just that. Small SMT cap with leads on the ends dipped in whatever epoxy-looking case they use.
At What Point is SMT Worthwhile?
- Thread starter thermionic
- Start date
Help Support GroupDIY Audio Forum:
I cut open a small radial cap once - something like a ceramic type - and was surprised to see just that. Small SMT cap with leads on the ends dipped in whatever epoxy-looking case they use.
ruffrecords
Well-known member
This is a fascinating thread. I set up my first SMT production line in the mid 80s. I was working for a technology consultancy and its was built as much to show customers what this new fangled SMT stuff could do as it was for short run production. Most of the products I designed throughout my working life used a mixture of SMT and through hole parts. When I retired and got back into tube audio it never occurred to me to use SMT and I think it is still hard to get high voltage caps in SMT not to mention tube sockets! However, I am pleasantly surprised to see that hand assembly is a practical proposition. My 65 year old hands are a bit shaky but but this thread has inspired me to give it a go. Could be a boon for EQ. Thanks guys for turning me on to SMT again.
Cheers
Ian
Cheers
Ian
People are often afraid by the small size of SMD part, but the truth is that with a minimal equipment (magnifier, tweezer), it is easier and faster to hand work SMD parts. The only point where you need specialized equipment is BGA parts where a reflow oven is mandatory (but it is not common to use BGA in audio circuits ;D ), and packages with thermal pads which need at least hot-air and a good practice. QFN can be touchy but can be done at home with some practice.
One point that is for me important is that it is easier to debug or modify a pcb populated with SMD passive part than TH. Changing a TH resistor or cap on a double side PCB can really be a PITA, with a non negligible risk of damaging the PCB, whereas it takes a few seconds (if you have a second iron it is really easy) to remove a SMD two terminal part without any risks.
It is true that high voltage caps (and you also have to be suspicious for resistor) are not really available, but there is no problems in mixing TH and SMD ... it increase the cost for "mass production" but for DIY this is not really an issue.
I am still using TH parts for electrolytics (more choice, same size) and film caps, connectors, but for resistors, transistors, ceramic caps, I go for SMD.
And Ian, I would like to thank you since you give me the idea of using those PCB sockets as tube sockets
One point that is for me important is that it is easier to debug or modify a pcb populated with SMD passive part than TH. Changing a TH resistor or cap on a double side PCB can really be a PITA, with a non negligible risk of damaging the PCB, whereas it takes a few seconds (if you have a second iron it is really easy) to remove a SMD two terminal part without any risks.
It is true that high voltage caps (and you also have to be suspicious for resistor) are not really available, but there is no problems in mixing TH and SMD ... it increase the cost for "mass production" but for DIY this is not really an issue.
I am still using TH parts for electrolytics (more choice, same size) and film caps, connectors, but for resistors, transistors, ceramic caps, I go for SMD.
And Ian, I would like to thank you since you give me the idea of using those PCB sockets as tube sockets
thermionic
Well-known member
- Joined
- Jun 3, 2004
- Messages
- 1,671
A concept that I find quite interesting is, if you have a machine - in a manufacturing scenario like Gareth - the machine effectively takes the place of what would've been drawers of components.
Am I correct to assume that a machine like Gareth's will know stock levels automatically? If so, then I'm also going to assume it integrates with inventory software... You could practically automate it to a degree where the machine almost files an order at Mouser, right?
If I were to go down the SMT route I think I'd opt for Chris' hybrid approach, where the laborious task of everyday passives is handled by the machine, but some of the more esoteric items such as precision caps are fitted TH by hand.
I got my first exposure to SMT in 1990, at a computer hardware design firm. Even then, some parts were only available in SMT packages.
Thanks again to all - fascinating thread.
edit - I think it is safe to assume that Gareth has inventory records at the click of a mouse. Even if the machine doesn't measure the quantity in its hopper, as long as you tell it you put in, say, 500 x 100R resistors, it's trivial for it to deduct one every time it pops one into a board.
Am I correct to assume that a machine like Gareth's will know stock levels automatically? If so, then I'm also going to assume it integrates with inventory software... You could practically automate it to a degree where the machine almost files an order at Mouser, right?
If I were to go down the SMT route I think I'd opt for Chris' hybrid approach, where the laborious task of everyday passives is handled by the machine, but some of the more esoteric items such as precision caps are fitted TH by hand.
I got my first exposure to SMT in 1990, at a computer hardware design firm. Even then, some parts were only available in SMT packages.
Thanks again to all - fascinating thread.
edit - I think it is safe to assume that Gareth has inventory records at the click of a mouse. Even if the machine doesn't measure the quantity in its hopper, as long as you tell it you put in, say, 500 x 100R resistors, it's trivial for it to deduct one every time it pops one into a board.
Why do you want a P&P machine ? Do you plan to have a serious production ?
For me, there are different scenarios :
- proto boards. Hand placement and soldering like described by Andy Peters (post #3)
- Small batch (10 boards or so). Stencil, solder paste, hand placement, small oven (this can be a DIY hack of a toaster oven)
- Production. P&P machine, which means maintenance, operation, and a large investment. At that point thinking of subcontracting PCB population (at least for SMD part only) is a good idea and may be more profitable than buying a machine.
For me, there are different scenarios :
- proto boards. Hand placement and soldering like described by Andy Peters (post #3)
- Small batch (10 boards or so). Stencil, solder paste, hand placement, small oven (this can be a DIY hack of a toaster oven)
- Production. P&P machine, which means maintenance, operation, and a large investment. At that point thinking of subcontracting PCB population (at least for SMD part only) is a good idea and may be more profitable than buying a machine.
thermionic
Well-known member
- Joined
- Jun 3, 2004
- Messages
- 1,671
Yesterday, I spent a lot of time populating the passives on the 390-part PCB. I know firms that will do this for me, but it will still take them plenty of time in a TH format, hence my asking 'at what point?'
I m not sure if it will ever have the volume to justify full SMT conversion, but at the same time, manually placing, soldering and cropping TH parts isn't fun.
I've had to stock up on DIP8 ICs due to TI culling their TH ranges. This is a worry.
I m not sure if it will ever have the volume to justify full SMT conversion, but at the same time, manually placing, soldering and cropping TH parts isn't fun.
I've had to stock up on DIP8 ICs due to TI culling their TH ranges. This is a worry.
I order PCB for SMD prototypes because you can't rats nest a prototype circuit together like the good old days with TH parts.dfuruta said:
The good news for modest sized PCB you can get small runs of boards dirt cheap.
JR
Andy Peters
Well-known member
dfuruta said:
PCBs all the way. When you can get ten 100 mm x 120 mm two-layer PCBs for $40 shipped, there's no point to dicking around with breadboards.
ruffrecords
Well-known member
Chris_V said:And Ian, I would like to thank you since you give me the idea of using those PCB sockets as tube sockets
They are pretty neat aren't they. I cannot take credit for them, I fist saw then in someone else's post here on groupDIY.
Cheers
Ian
Interesting. Any reference to a parts no. for those tubepin connectors?
Jakob E.
Jakob E.
Gareth Connor
Well-known member
Mmmm..... quite a lot to reply to......
Thermionic.
My old and trusty CS3 machine does not have parts inventory built in. The facility can be retrofitted. Some of my component feeders are from later-model machines that had the "smart feeder" facility. It would be convenient to have automated inventory and parts location identification, but for me it's not a show-stopper. The extra time involved in manual inventory maintenance is a small price to pay when offset against the time saved by having the P&P available.
Chris_V.
Why do you want a P&P machine ? Do you plan to have a serious production ?
Good questions!! I sometimes ask them of myself.....
The time consumed on hand placing, soldering, cropping of components was a factor. Bear in mind that I do this for a living and the quicker I can design and build a piece of equipment, the quicker I can move on to the next project, hopefully, having made a profit on the previous project.
Maintenance - part of the cost of ownership. There's first-line maintenance such as feeder cleaning & servicing, machine cleaning, alignment checks, etc., that are ideal to do in quiet periods and cost little more than cleaning & lubrication materials and time... Then there's the dreaded machine-failure maintenance which carries a heavy cost; the choice is to accept this as a possibility and plan for it or don't have the P&P facility.
Operation - a learning curve that is steep at first, but with practice, becomes part of the normal skills. Whilst the detailed skills are specific to a machine type, the overall principals are the same for virtually any P&P facility.
Large investment - Let's move on, shall we? Seriously, though, yes it has been a SERIOUS investment, and with hind-sight I may have decided differently. The P&P machine I bought was for a very good price and this was a major factor in setting-up the facility. Having the service back-up of a really good production machinery engineer was another major factor. However, it's not just the P&P machine & it's specialist service, it's all the other hardware that goes with it.... feeders of many types and quantities, specialised air compressor & air cooler, air-line installation (I'm good with copper piping), reflow oven, paste printer, space to fit it all in, uninterruptible power supply for the P&P (personal choice, not mandatory). When all this is weighed against the convenience of having the P&P on site and ready to use at a moment's notice, as well as being able to offer an assembly service as an add-on to design services, it can make sense. Some projects would not have been possibly without SMD and many would not have been cost effective without in-house P&P - remember that much of the audio industry is hell-bent on paying the lowest price under any circumstances. In a small and skilled business where there are a few equally skilled guys, 10 hours of manual assembly time costs the same as 10 hours of P&P time but the throughput of work is radically different, giving the P&P guys an edge.
JR.
How to cover the back side of the board with SMT. With the facilities that I have, hand-pop would be the only way for a board such as the one I showed. The use of glue to hold the bottom-side parts in place could be a method. There are other ways, but I expect the discussion would really warrant a separate thread. I'm happy to discuss if there is interest.
Land-fill with wire leads..... wow, that's a LOT of offcuts!
Wave.
0402. Respect, Sir!
Dfuruta.
Prototyping.
A combination of methods, depending on the end-product. The SMD version is always a PCB, but for me the lead-up to it can take 2 paths, sometimes a combination of both,
1. An all-theoretical path without any prototype PCBs prior to production.
This has recently been done on a couple of really complex analogue designs using 4 and 6 layer PCBs. Having the strength-of-character to go straight from a paper design to 6 layers is a moment of taking a very deep breath! I will only do this if the product is made-up from pre-known and proven blocks of circuitry, and after 33+ years there are plenty of these in the design arsenal. The skill is in tying all blocks together correctly - levels, polarity (phase), voltages, power supply, layer assignment, trace routing, etc.
2. Proto-board.
If I need a new circuit, this is developed on protoboard, the proven circuit is then drawn and added to the target design. As an old analog-dog, I'm not having to deal with SMD-only processor chips, thereby making protoboard viable.
An aid to timely turn-round of of custom designs has been the development of a range of very useful building blocks ( VUBBs ) such as input stages, output stages, buffers, VCA blocks, interlocking switch banks, relay switchers, FET switchers, CMOS switchers, meter drivers..... the list goes on..... http://www.cmlaudio.com/modular-assemblies/sub-assemblies/ All of these are on small PCBs that can easily be tied together to form larger systems. A new development such as a compressor would comprise of a set of building-block boards and a component-level sidechain, all built on protoboard, but in a fraction of the time that an all-discrete component design would take.
Thermionic.
My old and trusty CS3 machine does not have parts inventory built in. The facility can be retrofitted. Some of my component feeders are from later-model machines that had the "smart feeder" facility. It would be convenient to have automated inventory and parts location identification, but for me it's not a show-stopper. The extra time involved in manual inventory maintenance is a small price to pay when offset against the time saved by having the P&P available.
Chris_V.
Why do you want a P&P machine ? Do you plan to have a serious production ?
Good questions!!
I sometimes ask them of myself.....The time consumed on hand placing, soldering, cropping of components was a factor. Bear in mind that I do this for a living and the quicker I can design and build a piece of equipment, the quicker I can move on to the next project, hopefully, having made a profit on the previous project.
Maintenance - part of the cost of ownership. There's first-line maintenance such as feeder cleaning & servicing, machine cleaning, alignment checks, etc., that are ideal to do in quiet periods and cost little more than cleaning & lubrication materials and time... Then there's the dreaded machine-failure maintenance which carries a heavy cost; the choice is to accept this as a possibility and plan for it or don't have the P&P facility.
Operation - a learning curve that is steep at first, but with practice, becomes part of the normal skills. Whilst the detailed skills are specific to a machine type, the overall principals are the same for virtually any P&P facility.
Large investment - Let's move on, shall we? Seriously, though, yes it has been a SERIOUS investment, and with hind-sight I may have decided differently. The P&P machine I bought was for a very good price and this was a major factor in setting-up the facility. Having the service back-up of a really good production machinery engineer was another major factor. However, it's not just the P&P machine & it's specialist service, it's all the other hardware that goes with it.... feeders of many types and quantities, specialised air compressor & air cooler, air-line installation (I'm good with copper piping), reflow oven, paste printer, space to fit it all in, uninterruptible power supply for the P&P (personal choice, not mandatory). When all this is weighed against the convenience of having the P&P on site and ready to use at a moment's notice, as well as being able to offer an assembly service as an add-on to design services, it can make sense. Some projects would not have been possibly without SMD and many would not have been cost effective without in-house P&P - remember that much of the audio industry is hell-bent on paying the lowest price under any circumstances. In a small and skilled business where there are a few equally skilled guys, 10 hours of manual assembly time costs the same as 10 hours of P&P time but the throughput of work is radically different, giving the P&P guys an edge.
JR.
How to cover the back side of the board with SMT. With the facilities that I have, hand-pop would be the only way for a board such as the one I showed. The use of glue to hold the bottom-side parts in place could be a method. There are other ways, but I expect the discussion would really warrant a separate thread. I'm happy to discuss if there is interest.
Land-fill with wire leads..... wow, that's a LOT of offcuts!
Wave.
0402. Respect, Sir!
Dfuruta.
Prototyping.
A combination of methods, depending on the end-product. The SMD version is always a PCB, but for me the lead-up to it can take 2 paths, sometimes a combination of both,
1. An all-theoretical path without any prototype PCBs prior to production.
This has recently been done on a couple of really complex analogue designs using 4 and 6 layer PCBs. Having the strength-of-character to go straight from a paper design to 6 layers is a moment of taking a very deep breath! I will only do this if the product is made-up from pre-known and proven blocks of circuitry, and after 33+ years there are plenty of these in the design arsenal. The skill is in tying all blocks together correctly - levels, polarity (phase), voltages, power supply, layer assignment, trace routing, etc.
2. Proto-board.
If I need a new circuit, this is developed on protoboard, the proven circuit is then drawn and added to the target design. As an old analog-dog, I'm not having to deal with SMD-only processor chips, thereby making protoboard viable.
An aid to timely turn-round of of custom designs has been the development of a range of very useful building blocks ( VUBBs
) such as input stages, output stages, buffers, VCA blocks, interlocking switch banks, relay switchers, FET switchers, CMOS switchers, meter drivers..... the list goes on..... http://www.cmlaudio.com/modular-assemblies/sub-assemblies/ All of these are on small PCBs that can easily be tied together to form larger systems. A new development such as a compressor would comprise of a set of building-block boards and a component-level sidechain, all built on protoboard, but in a fraction of the time that an all-discrete component design would take.
ruffrecords
Well-known member
gyraf said:
Farnell #149319
http://uk.farnell.com/harwin/h3161-05/socket-pcb-gold-knurled-1mm/dp/149319?ost=149319&selectedCategoryId=&categoryNameResp=All%2BCategories
Cheer
Ian
For prototyping SMDs you can DIY etch one offs 8)
Flex style
Solid ink printer + Pyralux
Flex style
Solid ink printer + Pyralux
i'm a huge fan. even for prototypes. you need to have a look at what the Synth Module DIY guys are doing (over on forums like muffwiggler). There is zero fear of SMD there.
Monte McGuire
Well-known member
- Joined
- Jul 26, 2013
- Messages
- 384
For hand etched prototypes, SMD is easier than TH because you don't have to drill any holes! Granted, with only one side, you need to find some SMD jumpers to bridge traces, and still, your layout options are pretty limited, but it's nice to be able to go from layout to finished board in only a handful of hours.
Similar threads
