KMV254c: Gefell MV691 to KM54c modification.

[erikrex]

Thanks, @SparkleBear. I’m in for two PSU boards in addition to the mic boards.

Getting a bit busy this week, will catch up on your recent updates this weekend.

Cheers

 

[SparkleBear]

What is the idea behind the series connection of capacitors C52 and C56? Makes no sense to me atm.

This was a recommendation from Robert at RMS audio. I respect the man so I tried it out.

Wouldn't it be better to position R50 between C50 and C54?

Whats the benefit / difference the way it is vs the way you recommend?

I also put the .1uf caps closer to the H+ regulator as recommended.




Unless anything else is glaring, i will make an order for these pcb later this weekend.

 

[rock soderstrom]

This was a recommendation from Robert at RMS audio. I respect the man so I tried it out.

The two capacitors (with their parallel small foils caps) are in series which has the following consequences that the total capacitance is less than any one of the series capacitors’ individual capacitances!

https://www.allaboutcircuits.com/textbook/direct-current/chpt-13/series-and-parallel-capacitors/
This turns a 22uf and a 33uf capacitor into a resulting capacitor of 13.2uF!

The result is that you have less filter effect, you need more space on the PCB and have to spend more money on it. Not a good deal.

This kind of series connection of capacitors is done, for example, when the voltage capability of a single capacitor is not sufficient, as in old Fender guitar amplifiers. But that is not the case here, so there are no advantages IMHO, only disadvantages as described above.

Possible solutions would be:

- you use only one capacitor and save space and money with a better filter performance than before

- you continue to use two capacitors as before and connect a resistor in between and get even more filtering because you then have two RC filters after the LM317. Whether this is really necessary is another question, just like C55/C57, whose function is already taken over by C63 in my opinion.

BTW, it is better to make C56 and C62 identical with one capacity, that makes shopping a bit easier. Use 22 or 33uf for both if you want to keep the second cap.

Whats the benefit / difference the way it is vs the way you recommend?

The situation after the rectifier for B+ is similar. At the moment you do not have a smoothing capacitor directly after the rectifier, but two parallel capacitors after the first series resistor R50.

If you would do it like this --> rectifier - C50 - R50 - C54, your filtering would be better than it is now, with the same amount of components.

https://www.electronics-tutorials.ws/diode/diode_6.html
Just my 0,02$

Cheers!

 

[SparkleBear]

Thanks Rock! I think I will make those changes... That all makes sense. It is very quiet as it is right now though ;P

Ok, so I just got the PCBs and not the best news... Something got crossed with one pad's nets in Kicad and this happened.



The ground pad for the tuchel connector is just floating there. Needed to bodge it to ground plane. *FACEPALM* I will have to make that correction and print a new batch of PCBs. DOH... expensive mistake. Stay tuned.

 

[SparkleBear]

Good news is that is the only error on the boards. I built on this board and its working (with the bodged ground pad). I took this opportunity to make some final adjustments to the board, including a couple silk improvements and added another jumper to configure the cathode resistors. Widened the holes for 100k and cathode resistors incase 1/2w is used. I also made the wire entry pads a bit bigger to accommodate the nice hookup wires chunger sells on his site. (my favorite). Oh, and most importantly, I rounded the PCB corners a bit more. ;P

Definitely a bummer to waste a production of boards but I want this to be turnkey for y'all. Going to order another batch this weekend. And I will also post a little step by step build documentation from the mic I built yesterday.

 

[elhorst]

two mic pcbs and a psu pcb please...

 

[SparkleBear]

Hang in there y’all, it’s so close… ordering new pcbs for the mic and psu today. I’ll be out of town for a week or two. And whenever I’m back I can check over the new boards and if they are all good I can send them out.

Thanks for the interest and support.

 

[SparkleBear]

The two capacitors (with their parallel small foils caps) are in series which has the following consequences that the total capacitance is less than any one of the series capacitors’ individual capacitances!

https://www.allaboutcircuits.com/textbook/direct-current/chpt-13/series-and-parallel-capacitors/
This turns a 22uf and a 33uf capacitor into a resulting capacitor of 13.2uF!

The result is that you have less filter effect, you need more space on the PCB and have to spend more money on it. Not a good deal.

This kind of series connection of capacitors is done, for example, when the voltage capability of a single capacitor is not sufficient, as in old Fender guitar amplifiers. But that is not the case here, so there are no advantages IMHO, only disadvantages as described above.

Possible solutions would be:

- you use only one capacitor and save space and money with a better filter performance than before

- you continue to use two capacitors as before and connect a resistor in between and get even more filtering because you then have two RC filters after the LM317. Whether this is really necessary is another question, just like C55/C57, whose function is already taken over by C63 in my opinion.

BTW, it is better to make C56 and C62 identical with one capacity, that makes shopping a bit easier. Use 22 or 33uf for both if you want to keep the second cap.



The situation after the rectifier for B+ is similar. At the moment you do not have a smoothing capacitor directly after the rectifier, but two parallel capacitors after the first series resistor R50.

If you would do it like this --> rectifier - C50 - R50 - C54, your filtering would be better than it is now, with the same amount of components.

https://www.electronics-tutorials.ws/diode/diode_6.html
Just my 0,02$

Cheers!

Made adjustments to the PSU. What do you think now?





Removed the two small value bypass caps from c56 and c52. Made them both 22uf (rather than 33uf). Added a 47uf smoothing cap after heater rectifier bridge. Put r60 after c50 rather than right after the LM317 output...

Feeling good about it. Unless I get some constructive feedback about this revision I will send it to the fabricator before monday.

Thanks for the suggestions Rock.

 

[Khron]

Any reason why the ground is not properly earthed?

 

[rock soderstrom]

Made adjustments to the PSU. What do you think now?

I think there is still some work to be done. I have briefly looked at the real implementation on the PCB, some questions arise for me:

What is the distance between the high-voltage traces to each other and to Ground?

I would create more distance because I would be afraid of flashovers.

Only as an example this place (there are still further). Here the two traces come very close to each other without need....



What is the AC voltage of the secondary side of the planned transformer for the heater? Really 20V AC?

Is this AC ground thing really safe? I have no experience with it...

PS: C53/C55 one of the two is superfluous from my point of view. The same applies to C36/C33.

 

[Khron]

What is the distance between the high-voltage traces to each other and to Ground?

I would create more distance because I would be afraid of flashovers.

That crossed my mind the first time i saw the PSU board, but i didn't want to help with ALL of the design of this thing...

Is this AC ground thing really safe? I have no experience with it...

https://en.wikipedia.org/wiki/Ground_(electricity)
Quite the contrary, the entire point is *TO* be safe. And also to help minimize hum, as @Spencerleehorton has recently found.

https://groupdiy.com/threads/m49-oliver-archut-alt-version.79640/page-4

 

[rock soderstrom]

Let's talk about fuses for a minute. It seems that only the primary side of the transformer is fused. I wouldn't do this on the PCB but mounted on the housing and accessible from the outside.
As it is, someone inexperienced will certainly connect both mains conductors here and generate a short circuit.



I may be a bit over-cautious about fuses, but for this application I would fuse both secondary windings as well.

I have seen a unprotected transformer burn...

Similar example:



Again, just my 0,02$. It is just the way I do it.

 

[SparkleBear]

I think there is still some work to be done. I have briefly looked at the real implementation on the PCB, some questions arise for me:

What is the distance between the high-voltage traces to each other and to Ground?

I would create more distance because I would be afraid of flashovers

Good catch on that. I think the earlier revisions were spacious like this... but i nudged things around and didnt clean the traces. Thanks rock

Only as an example this place (there are still further). Here the two traces come very close to each other without need....

View attachment 111199

What is the AC voltage of the secondary side of the planned transformer for the heater? Really 20V AC?

No, its actually 12v. The first chinese transformer i was working with was labeled 200v and 20v *shrug*

Is this AC ground thing really safe? I have no experience with it...

This is supposed to reduce hum... i wanted to possibly use this psu for a schoeps m221b that I was messing around with...

PS: C53/C55 one of the two is superfluous from my point of view. The same applies to C36/C33.

Removed the .0047uf caps.

Let's talk about fuses for a minute. It seems that only the primary side of the transformer is fused. I wouldn't do this on the board but mounted on the housing and accessible from the outside.

I build with the fuses on either the IEC plug or the blade fuse will be transfered from the donor alctron pcb. That fuse is on the hot side going to the power switch... Then the switch feeds the transformer directly... Then the secondaries are connected to hv and lv... are you saying I should also put a fuse at the hv/lv pairs? Right before the rectifiers?

As it is, someone inexperienced will certainly connect both mains conductors here and generate a short circuit.

I may be a bit over-cautious about fuses, but for this application I would fuse both secondary windings as well.

I have seen a unprotected transformer burn...

 

[Murdock]

This whole grounding thing is really hard to grasp...
If having B+ 0V and H+ 0V seperated and only connected at the output connector would it be advisable to have two ground planes next to each other or are traces better in that case?
Is it even needed to have them seperated and is it better to have a big copper plane for the whole 0V?

Maybe this is a topic for another thread but I thougt as SparkleBear and Rock are both designing their own PSU PCB's it would fit in here.

 

[rmburrow]

Hang in there y’all, it’s so close… ordering new pcbs for the mic and psu today. I’ll be out of town for a week or two. And whenever I’m back I can check over the new boards and if they are all good I can send them out.

Thanks for the interest and support.

Will these boards fit inside a Schoeps M221a? What is the cost of the mic board?

 

[SparkleBear]

I think the m221b

Will these boards fit inside a Schoeps M221a? What is the cost of the mic board?

The m221 is a 20mm microphone. The MV69x is 25mm. Also, the one side of the capsule is directly grounded on the schoeps. That is not how the scheme goes on the original km54 and in turn the modded kmv254c (my board).

I did design an m221b board that fits into the 20mm MXL bodies but i never made a PSU that was quiet enough for that cathode scheme.

 

[rock soderstrom]

Good catch on that. I think the earlier revisions were spacious like this... but i nudged things around and didnt clean the traces. Thanks rock

But this is not the only place where the distances between traces with high voltage are really small. This is also due to the groundplane (and the corresponding design rules) in the B+ area. I would research again what the minimum distances are for voltages of this type. There are certainly rules for that.

No, its actually 12v. The first chinese transformer i was working with was labeled 200v and 20v *shrug*

12V is quite a lot, 9V would be better from my point of view. Why?

12VAC gives about 17VDC after rectification. If we subtract the 1.5V for the diode losses and 0.5V for the series resistor and the choke, we are left with 15V DC at the input of the LM317. This means the regulator converts 8.7V (with a 6.3V/0.175mA heater) into heat.

8.7V x 0.175A = 1.52W ---> this is not little and can cause problems for the LM317 depending on the heat sink and ambient temperature (e.g. in a closed case after 14h of recording!). In reality, it could be even more, as the mains transformer may not be fully loaded and supply slightly more voltage.

If you also want to operate other tubes with maybe 300mA heater at 6.3V with this power supply, the power loss (heat!) almost doubles with all consequences.

Just as a thought experiment...

This is supposed to reduce hum... i wanted to possibly use this psu for a schoeps m221b that I was messing around with...

I would read up on this topic again. Safety is king, so be sure to read member @Khron contribution.

I would also not ground directly after the rectifier but at the respective output of the two power supplies. Otherwise there may be trouble with humming.

I build with the fuses on either the IEC plug or the blade fuse will be transfered from the donor alctron pcb. That fuse is on the hot side going to the power switch... Then the switch feeds the transformer directly... Then the secondaries are connected to hv and lv... are you saying I should also put a fuse at the hv/lv pairs? Right before the rectifiers?

As I said, I am not a big fan of primary fuses on the PCB as it is at the moment. Others do it, but I wouldn't do it. (-->safety and "HF dirt from the outside danger" ;-)

Yes, I would fuse both PSUs (B+/F+) separately, just before the rectifier. See my design as an example.

This whole grounding thing is really hard to grasp...
If having B+ 0V and H+ 0V seperated and only connected at the output connector would it be advisable to have two ground planes next to each other or are traces better in that case?
Is it even needed to have them seperated and is it better to have a big copper plane for the whole 0V?

I think it is better to keep the two grounds separate, even if you then combine them at the exit. There may be reasons not to do it that way. An example would be negative heater voltage due to cathode bias generation as with the U67 or elevating the heater due to limited Ufk.

I can't say whether groundplanes or traces are better, but ideally they should be the same, if you don't consider the production processes and possible savings there.

In my example I have a bridge on the PCB which can be easily removed if necessary! However, the grounds are built separately.

 

[SparkleBear]

Hey all, Back from a little road trip and made some more refinements of the psu based on advice from y'all...

Split the ground planes and made a jumper scheme to connect them up as needed. Added PTC after the rectifiers on each rail... increased clearance of copper pour on B+ gnd plane.




Thanks for the advice!

-Eric

 

[SparkleBear]

Hey y'all, final boards for the mic and psu are both heading to me from the fab this next week...

I have completed a current BOM and generated an interactive BOM to help with assembly too.

Here are the mouser carts:

UVMPSU: https://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=6710cb17f7
KMV254c: https://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=aaaa414db9

BOM for PSU and MIC: (check attached zip file)

Stay tuned for build thread

 

[rmburrow]

I think the m221b

The m221 is a 20mm microphone. The MV69x is 25mm. Also, the one side of the capsule is directly grounded on the schoeps. That is not how the scheme goes on the original km54 and in turn the modded kmv254c (my board).

I did design an m221b board that fits into the 20mm MXL bodies but i never made a PSU that was quiet enough for that cathode scheme.

Clone the Neumann NKM supply; that works with the Schoeps M221 mic nicely. Do you have any m221B boards? I can modify them for the M221a.