FEELER - Universal Tube Microphone PSU

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
It is designed to fit into the acatel  (sp?) enclosure and as long as the 49's voltage requirements are within the range specified by Matador for the psu you should be fine
 
OK. . . matador still does not have a proper workspace set up, so it is incumbent upon me to attempt to prototype the 2nd revision prototype PCB.  I am hoping I do not create too much smoke and fire!  Matador has insisted I post photos for maximum entertainment value to the prototyping dollar.  So, here we go.

p99507341-4.jpg


I was given a build sequence, so I begin with the first stage:

First Stage:
1) Stuff D3, D5, D6, D7, D9, D10 (1N4007)
2) Stuff D1 (1N5380)
3) Stuff R1 and R3
4) Stuff C15 and C1, C2
5) Stuff and heatsink U1

p297318978-4.jpg


With the components for the first stage in place, I prepare to power up the board.  Yes, the fuse is precariously hanging from the assembly, but I figured this board is going to go in and out of the case many times before I am finished, so I opt for a non-permanent solution.

p357986493-4.jpg


If all went well, I should see a fixed 225V  +- 1% at the node labelled "VFIXED".

No smoke, but, all does not go well. . . in fact, I see nothing except directly at the transformer secondaries where I get 230V. . .. now that seems quite a bit out of spec  :eek: .  Gotta love Chinese quality control!

p239887582-4.jpg


And here is my "VFIXED" voltage. . . or not voltage.

p493480223-4.jpg


I scratch my head and wonder about it for a while. . . then, I call Matador and get voicemail. . . search the internet for full wave rectifier diagrams and trace the board.  I probably should have asked for a schematic  ;) .  It occurs to me I do not see a ground anywhere, so the circuit is incomplete.  A text message from Matador confirms there was an error in the layout, and the schematic was not properly applied to the layout in software.  Matador provides a fix which I implement with a piece of 22AWG wire.

p78160899-4.jpg


OK. . . now the first stage is alive!  Still no smoke!

p373917644-4.jpg


"VFIXED" sits at 238V.  It should be 225V.

Voltage is not where it should be.  My Vreg across R1 is 1.32V when it should be 1.25V.  Not what we expected, but it will not blow up the 2nd stage, so I continue.  In theory, without making a resistor change, this PSU will only be able to go to a minimum of 115V, so no U47 unless I make a correction.

Next, I populate the 2nd stage of the B+ circuit.

Second Stage:
1) Stuff D4, D8 (1N4007)
2) Stuff D2 (1N5380)
3) Stuff R2 and R4
4) Stuff pot R5
5) Stuff C3
6) Stuff and heatsink U2

For the heatsink, I break out the thermal compound. . . i don't really anticipate trying to drop from 230V to 105V for U47 application in a fully enclosed, small case like this, but theoretically, this PSU should be able to do the job.  If it does, this 2nd stage regulator will be called about to do the lions share of the regulation and it will get hot.

First some compound directly on the heatsink.

p356678697-4.jpg


Then, some on the back of the regulator.

p297552139-4.jpg


And we squish it all together with the insulating strip between.

p22931905-4.jpg


2nd stage populated!

p282169199-4.jpg


This is designed to give us 85V-220V range . . . and no smoke or flames.  Let's see what we get.

The high side looks fine at 222.3V.

p254637386-4.jpg


And, in this configuration, our supply bottoms out a 126.5V.

p321840447-4.jpg


I figure it's pointless for me to stab and guess at resistor value changes for the 1st and/or second stage right now.  We will not blow anything up component-wise, so I figure I will wait for Matador to wake up and read this post and then recommend alt values.  For now, I continue to the fixed bias supply which is next in the chain.

Bias Supply

1) Stuff D18 through D22
2) Stuff R10 and R12
3) Stuff R11
4) Stuff C8 through C13
5) Stuff IC2

Measure BIAS+ - it should measure a fixed -2.5V (wrt. ground) regardless of the setting of R11
Measure BIAS- - it should measure between 0V and -2.5V (wrt. ground) by turning R11

p99051777-4.jpg


I do in fact get the expected -2.51V at BIAS+ and BIAS- allows adjustment between 0V and -2.51V!

p260309515-4.jpg


p230362509-4.jpg


Next, I install D17. . . and poof!  bias disappears. . .  :mad:

In his notes, Matador indicated that he was not sure if the polarity was printed correctly for D17.

p11789143-4.jpg


It seems like I use the Hakko 808 desoldering tool at least once on every project!

p512461465-4.jpg


The 808 makes quick work of removing and flipping D17.

p137248922-4.jpg


With D17 installed, I now get -7.8V at BIAS+ and on BIAS-, can adjust between -7.8V and 0V.

p54243826-4.jpg


I set my bias to -1V since that seems a reasonable starting point and leave it there for now.

p6174675-4.jpg


That's all 'til next time.  Hopefully Matador can provide some alt values, and I can go to the local electronics store and pick up a couple resistors tomorrow and then continue.
 
Hey Chung.  I think those transformers were rated on less than 120VAC and will therefore output 230-240 consistently if labeled 200.  The ones labeled 175 actually output 200V or close to it. 

At least that is the case with the few I've seen. Maybe you can test this further.    They may be using a reference AC voltage of 110 or 115 when we genuinely have 120 at the outlet here in the U.S. 
 
Argh, TI specifies 1.3V MAX on VREF (1.27 typical), and we are getting 1.322V???  Unacceptable TI!.

My paper napkin math says for that VREF, the 39.2K resistor needs to be dropped to 37.4K to drop the output down to 225V.  I would also pick up a 33K (or 33.2K for 1% value) so we can try a fixed 200V supply.

Category 5 said:
Hey Chung.  I think those transformers were rated on less than 120VAC and will therefore output 230-240 consistently if labeled 200.  The ones labeled 175 actually output 200V or close to it. 

At least that is the case with the few I've seen. Maybe you can test this further.    They may be using a reference AC voltage of 110 or 115 when we genuinely have 120 at the outlet here in the U.S. 

I had him measure the DC voltage after the rectifier diodes (right at the first filter cap), and he gets 298V, which is consistent with a 200VAC voltage out of the transformer (should be 282V at nominal supply).  However it wouldn't surprise me if a) some were mislabeled, or b) they have ridiculous winding differences from unit to unit.

Chunger, I would clip out D17 and not worry about it for now.  The fact that you read -7V at BIAS+ means the diode is bypassing the reg.  BIAS+ should sit at -2.5V forever (wrt. ground that is).  D17 is actually not useful because the reg's output is always forced to ground.
 
This is looking really nice guys.

BTW - Everyone should hit up ebay for a Hakko 808 de-soldering tool.  Since the newer yellow and blue version came out the originals are going pretty cheap, and they will literally change your life the next time you have to remove a component from a through hole PCB.  Not only does it come off quick and painlessly in seconds, you can almost always re-use the component too.
 
Category 5 said:
This is looking really nice guys.

BTW - Everyone should hit up ebay for a Hakko 808 de-soldering tool.  Since the newer yellow and blue version came out the originals are going pretty cheap, and they will literally change your life the next time you have to remove a component from a through hole PCB.  Not only does it come off quick and painlessly in seconds, you can almost always re-use the component too.

Absolutely a must-have tool.  Why spend hundreds sometimes thousands of dollars on DIY kits and parts and risk damaging components and PCB's?

If you have a good large-chain electronics store like Fry's Electronics local to me, you may want to check pricing.  I have found Hakko products including the 808 at significant lower prices than online.
 
OK. . . back to it.  D17 coming out.  Matador informs me that the orientation is actually labelled correctly.  Some more testing on that section of the circuit may be in order, but for now, no ill effects on the bias circuit with it removed.  I cut instead of de-solder because I plan to clean the back side of the PCB with isopropyl alcohol and do not want it seeping to the front side where it is harder to scrub :p

p289457268-4.jpg


Next, the 39.2K R3 resistor comes out.  Hakko 808 making a 2nd appearance.

p43144740-4.jpg


Per matador's recommendation, I replace with 37.4K resistor.

p288991746-4.jpg


OK. . . now to power it up again and see what we have.  VRECT is sitting at 280.5V

p94533760-4.jpg


VREG is sitting at 215.1V . .  . slightly off our 225V target.

p490442988-4.jpg


Our reference voltage across R1 is now 1.253V

p57591523-4.jpg


OK. . . our minimum B+ is sitting at 123.2V . .  . so still far from our 105V target to get down to U47 application, and our max B+ value is 211.6V which is good enough to power U67.

p198299201-4.jpg


p247200515-4.jpg


I will wait for further diagnostics and advise about tweaking the 1st and 2nd stage B+ sections, but for now, I populate the remainder of the board to test the heater supply.

p108349758-4.jpg


Before I power up and test, I need to make a little jumper from a piece of cutoff.

p829798627-4.jpg


The heater supply is floating so we can easily configure it for positive or negative heater supply voltage.  If we ground the "H+" terminal on the terminal block with our jumper, we get a negative heater supply voltage from the "H-" terminal.  Conversely, if we ground the "H-" terminal with our jumper, we get a positive heater supply voltage from the "H+" terminal.

I start by testing the conventional positive heater supply like most microphones utilize.  So I jumper "H-" to "GRD".

p719667467-5.jpg


Right out of the gate, we have a 7.24V supply. . . and no smoke!  Excellent.

p715111043-4.jpg


We have a minimum of 1.251V and a maximum setting of 11.23V.

p679702540-4.jpg


p541116035-4.jpg


Next I put the jumper between "H+" and "GRD"

p605175652-4.jpg


Measuring the voltage at "H-", we see that we have a negative heater voltage adjustable in the same range.

p1031591228-4.jpg


Being as this supply will likely be tested on a C12 clone first, I set my jumper to ground "H-" and set heater supply to 6.3V.

p727684309-4.jpg


Getting close now. . . Hopefully we can crack the B+ adjustment range issue to hit our 210V-105V target.  Once that's done, we can torture test it on a U47 clone (or likely test resistor substitute) and see where our temps end up internally.
 
Ok, very close now.

VREF is specified at 15mA output current.  For the first reg, we have 225V/37.6K ~ 6mA (or 1.25/221 = 5.6mA, almost the same).  Second reg adds another 6mA, so the total from the first test is ~12mA.  So we'll need to retest everything once we have a dummy load in place.  I can adjust the resistors down to give the proper load if the voltages are fluctuating due to the minimum load.

First off, if the output cannot go below 126V it may be due to the Zener turning on.  Clip out D2 by cutting one side, then lifting it up slightly out of the way so it can't turn on.  We can push it back down and re-solder it in place later.  According to the datasheet, the minimum turn on voltage is 114V, so with 215V pre-reg it shouldn't turn on until B+ tried to go below 100V.

I'm guessing the real problem with the VFIXED at 215V and the output minimum at 126V is the minimum regulator load...so....

...let's establish a 20mA minimum load for both regs.  Place a 100V/20mA ~ 5K (4.7K works) load on B+ (between B+ and ground).  Power in the load is (20mA * 20mA) * 4.7K = 1.8W, so use appropriately beefy resistor. :)  Then if you could measure:

1) Vrect
2) Vfixed
3) VREF for both regs (for reg 1, measure across R1, and for reg2, measure across R2)
4) The output range by turning the R5 pot fully to both extremes

If the 20mA test load causes the voltages to fall in line then I have some resistor tweaks to make this work better without load.

After removing D17, does bias+ sit at -2.5V now?  What is the voltage of the net where R6 and C5 meet (or C5 cap + terminal)?
 
And hopefully everyone reading isn't bored to tears with this conversation - if nothing else, you can see what goes on to vet out these new designs.  ;D
 
Matador said:
And hopefully everyone reading isn't bored to tears with this conversation - if nothing else, you can see what goes on to vet out these new designs.  ;D

and people can better appreciate how great it is to get a fully tested,  dialled in and complete diy kit, one shipping charge, build it; done !!

really appreciate all your efforts  !!
 
Alright!  Sparks, smoke, and fun breaking parts, and other high-voltage related stuff this round!  All the things that make me glad I upgraded to a well-protected multimeter.

I didn't have a 4.7K or 5K higher-ish wattage resistor to test with, but I had a whole bunch of wirewound resistors left over from mk47 PSU builds that I ended up changing to chassis-mount type, so I figured I'd use them here.  I also had a "short" piece of G10 turret board material left over and hadn't used my hole punch in a while. . . so I thought i'd make myself a little test resistor board with 1K and 680R resistors.  Maybe I can use it to test other PSU's also.

p714575541-4.jpg


p641085281-4.jpg


p715848301-4.jpg


Next,  I remove D17 per Matador's recommendation.

p848675851-4.jpg


Then, I double check just to make sure things are behaving like they were before without load. . . and it does seem to be.

p980621834-4.jpg


My test load ends up being 4.612K . . . I figure that should be close enough to the 4.7K requested.  Guess I could have used 5K also since I put a few extra resistors on my little board.

p150455373-5.jpg


And, this is how I connected things up for some measurements.

p193707970-4.jpg


With load and nothing else changed, we are getting 103.4V-204.8V adjustment range now.

p54264483-4.jpg


p202345913-4.jpg


Next, I went to measure the VRECT, and this is where things started going south. . .

p375457660-4.jpg


I had a big SNAP!  and sparks near the node I was probing.

p361562426-4.jpg


You can see where as I was pressing the probe against the board, it shorted against the trace that was directly under the diode.  There is a crater and burn mark there where the soldermask has burned up and the trace is exposed.  A couple of pops, and I finally figured out what was causing it.  My brand new "Probemaster" brand  probes are magnificently sharp. .. . sharp enough to punch right through the soldermask.  So, I may not be able to take as many photos while making measurements on high voltage sections. . .

At this point, VFIXED is still sitting at 215V, so 1st stage is undamaged.

p139125429-4.jpg


reference voltage across R1 is 1.253V

p353794486-4.jpg


When I tried to measure reference voltage across R2, I shorted the circuit again. . . This time from B+ (one of the nodes being probed) to the ground plane.

p205158372-4.jpg


I texted Matador, and he indicated that the 2nd stage regulator was probably toast. . . PERFECT!. .  . that's the one with thermal compound on it that is nearly impossible to reach now that the heatsinks are soldered in.  The acrobatics required to change it out would have been extremely difficult if not impossible without a good de-soldering tool like the Hakko 808.

p429665635-4.jpg


After replacing the 2nd stage regulator, VRECT and VFIXED were measuring 242V. . .  this meant that the 1st stage was damaged as well.  I'm not exactly sure how the diode function on the multimeter operates, but it made prolonged and sustained beeping noises at me when I probed D1, and I was unsure about the reading I was getting on D3, so I decided to replace those. . . again, the position of these diodes on the finished PCB is ridiculous for replacement.  So. . . this is where Hakko 808 de-soldering tool REALLY saved me.

p437197151-5.jpg


For the sake of sanity, I decided to re-install on the back side of the PCB.

p49915221-4.jpg


Here's a photo of my zapped DMM tip.  It has dulled and the gold plating has been removed from the spots where we arced.  Bummer. . . I really liked that these probes only registered .1R self resistance. . . now they are sometimes .2R or .3R  :mad:

p69518216-4.jpg


OK. . . now that the PSU is operating normally again, here are the rest of the requested measurements.

VRECT = 242V
VFIXED = 214.6V
B+ = 104V-204.7V
1st stage reference voltage (across R1) = 1.249V
2nd stage reference voltage (across R2) = .668V-1.257V


 
You rock Chunger!

Everything appears to be in order now:  I have a small list of edits that i'll get into the next prototype (shouldn't take long).

One last thing to measure to understand the D17 issue...can you measure:

1) The DC voltage (wrt. ground) at the "In" terminal of IC2 (it's where R9, C10, and D17 meet)...or just the + side of C10 if that is easier
2) The DC voltage (wrt. ground) at the rectifier diodes (the net "in-between"D18 and D19, and again between D21 and D22)

It's possible that I have the polarities of C8 and C13 incorrect, which is causing strangeness.
 
Well, just a boring update. . . I blew up too many parts to continue on the one Mouser order, and Matador changed some components, so I need to gather myself a bit and "inventory" my stash for this project and place a new parts order before proceeding to the next prototype.

But, I figure I should get this current prototype finished out and ready to run on a C12 so that I can test and see how it sounds on an actual mic.

Being as Mouser is out of the Lorlin pattern switches, and I only have a handful left for kits in the webstore, I re-use the China special for this prototype.

p41373912-4.jpg


Since the PCB will not be coming in and out of the chassis regularly at this point, I screw in the fuse carrier.

p32290555-4.jpg


Chinese pattern switch returns to its rightful place.

p365919405-4.jpg


p518302718-4.jpg


The 7 pin XLR gets de-soldered and a new set of wires installed.

p114403000-4.jpg


And, I begin wiring up the PSU for operation.

p287568254-4.jpg


The finished prototype lays out quite nicely.

p398898610-4.jpg
 
Back
Top