Warm Audio WA-67 - Teardown

[Whoops]

You're right, you don't need to regulate the microphone voltage...but you seem determined with hitting 210V on the B+ for the microphone. Just suggesting how to do it with what you've got.

Yes, thanks. Those are good suggestions and I might implement them.

 

[TheJames]

(Nothing relevant here)

 

[Whoops]

I sent a mail to Warm Audio inquiring about the capsule.
I confronted them with the fact that the capsule used in this mic is not an exact K67 replica, with the same exact hole pattern and they brag in their website:

""we went to great lengths to recreate the capsule from the original ’67, making sure that ours performs the same way, has the same frequency response and contributes to the same chewy character as the original. That meant not only ensuring that the hole pattern was identical"

I got a first email saying they were checking it and would get back to me, but then I go no reply after 2 weeks.
Still waiting for their reply

 

[abbey road d enfer]

With 253 at the rectifier , 190 at the mic, or 210 at the mic once modified, and 400V caps, the zeners seem pretty much pointless.

I believe one of the possible explanations is that they are there to prevent excessive voltage building up when there is no load, which is the case in the time where the heaters have not yet reached their temperature, or when the mic is disconnected. The caps will stay charged at 250+ V for a long time. It may be a simple precaution for ensuring that the nominal voltage of the connectors is not exceeded.
UL/CSA and other agencies would frown at it.

 

[abbey road d enfer]

As for the transformer, I don't need to get a better one, I can just rewind this one if needed.

That won't answer for the possible core saturation.

For now I will just like to know how can I test or measure if the mains voltage is distorted or the transformer core is saturating?

You need an oscilloscope: just probe the low-volt secondary. The loss of about 10V on teh peaks should be easily visible, particularly if you compare with a good known xfmr.

 

[abbey road d enfer]

I'e tried R1 & R2 with 6.8k as you suggested but B+ was reaching 212V, and my goal was to have it around 210V. So I tried 9K resistors and that provides the 210V.

The difference between 210 and 212 is marginal in terms of operation, however, conceptually, there is a significant difference: in the latter case the zeners are doing their job (more or less, I admit), and in the former they don't.

But do we really need regulation here?

Maybe, if only for the reason I expose in post #245, which is probably not relevant to you, but is for a manufacturer. This is assuming the manufacturer knows what they're doing...

U67 PSU's are not regulated and that doesn't seem to be a problem.

At the time, Neumann did not have to face the numerous regulations we have today.

I know the Zeners are there with the intention of providing regulation at 212V, but how important is that?

I don't know, I can only guess (maybe wrongly )

, I may completely ... change their placement

I could easily do that if needed.

Not so easy because you would have to change the zeners for a higher value, roughly according to the voltages you would measure on the various caps., with the values of R1 and R2 changed.
For example, the one before last cap would see a voltage of about 217V, so you would need to increase the voltage of the zener string by the same amount.
You say you replaced them with 9k. Am I right in surmising they're actually 9.1k?

 

[Whoops]

That won't answer for the possible core saturation.

What could be the reason for the Core Saturation?
It's a reasonably sized transformer for the low current consumption on the high voltage rail, why is it saturating and how could it be solved?

You need an oscilloscope: just probe the low-volt secondary. The loss of about 10V on teh peaks should be easily visible, particularly if you compare with a good known xfmr.

I probed with the oscilloscope, did you mean High-Voltage secondary and not Low-Volt secondary?
is that the Low Volt secondary is for the 6,3V Heater supply

So checking with the Oscilloscope, between the high voltage secondary and the bridge rectifier this is what the scope gives:



If I disconnect the secondary from the PSU circuit and Scope it (just the transformer) I have a perfectly round and symetrical sinewave.

 

[Whoops]

The difference between 210 and 212 is marginal in terms of operation

yes, for sure. 210V or 212V will be marginally the same. If in the end 210V is possible to achieve easilly fine, if not between 212V to 208V is good

Maybe, if only for the reason I expose in post #245, which is probably not relevant to you, but is for a manufacturer. This is assuming the manufacturer knows what they're doing...

I don't think they know what they're doing to be honest, at least not with this PSU as we can attest.

Not so easy because you would have to change the zeners for a higher value, roughly according to the voltages you would measure on the various caps., with the values of R1 and R2 changed.
For example, the one before last cap would see a voltage of about 217V, so you would need to increase the voltage of the zener string by the same amount.

Sorry Abbey, I didn't understand this part, could you please explain again.
Are you saying the last cap, C6 in the schematic? So the string of Zeners will be between R5 and C6 in parallel with R18, wouldn't increase the voltage value of the Zeners make it regulate to an higher voltage?
What I understood was that if voltage hits the Zener higher than 212V then the voltage will be regulated to 212V.
Maybe I'm missing something here, thanks for explaining

You say you replaced them with 9k. Am I right in surmising they're actually 9.1k?

yes, they're 9.1K

 

[abbey road d enfer]

What could be the reason for the Core Saturation?
It's a reasonably sized transformer for the low current consumption on the high voltage rail, why is it saturating and how could it be solved?

All xfmrs saturate, more or less. Cheaper ones tend to saturate more than expensive ones. Note that I'm oversimplifying here, but te subject could take pages.

I probed with the oscilloscope, did you mean High-Voltage secondary and not Low-Volt secondary?

Core saturation is visible on all secondaries. Hoawever, the loss of peak voltage also depends on many factors and it would probably different on the different secondaries.

is that the Low Volt secondary is for the 6,3V Heater supply

Yes, I recommended that because there are les lethal voltages.

So checking with the Oscilloscope, between the high voltage secondary and the bridge rectifier this is what the scope gives:

View attachment 89335

If I disconnect the secondary from the PSU circuit and Scope it (just the transformer) I have a perfectly round and symetrical sinewave.

OK, so it looks like core saturation is not the main issue.
Actually I had not been thinking straight. Since there's a bridge rectifier, the secondary votage should be measured across the secondary, either with a floating oscilloscope, or with a dual-trace scope in differential mode.
I believe the problem is the DCR (and maybe leakage inductance) of the xfmr. Can you measure both primary and secondaries DCR?

 

[abbey road d enfer]

Are you saying the last cap, C6 in the schematic? So the string of Zeners will be between R5 and C6 in parallel with R18,

That's where they are ATM. What I suggest is connecting the zener string across C5. Due to the +/-7V drop across R5, in order to get 210 at B+, you need to have 217 at C5; the zener string should then be a nominal 217.
Same if you connected them across C4, you should adjust the value at about 224.

What are the voltages with R1=R2=9.1k?

 

[moamps]

Interestingly, the Neumann transformer in NU67 is also saturated, which may mean that this was done intentionally, in order to obtain better stabilization of the output voltage for mains voltage fluctuations.
Also interesting is the design of the heater power supply circuit, especially if you keep in mind that this voltage participates in the bias of the electron tube and should be very low noise. Perhaps the performance of power supply in the WA with the LM317 is worse in this regard.



An interesting (and somewhat funny) debate about NU67 can be found here:
https://repforums.prosoundweb.com/index.php?topic=37309.0

 

[Whoops]

I believe the problem is the DCR (and maybe leakage inductance) of the xfmr. Can you measure both primary and secondaries DCR?

for sure, I will measure it when I’m back at the workshop, thanks

As far as I understand a transformer core saturates when it’s not big enough for the application.
But in this case the transformer is not small and seems reasonably sized for the power consumption demandings of the microphone.

 

[abbey road d enfer]

for sure, I will measure it when I’m back at the workshop, thanks

As far as I understand a transformer core saturates when it’s not big enough for the application.
But in this case the transformer is not small and seems reasonably sized for the power consumption demandings of the microphone.

Actually, a very common problem is winders use charts to figure out the core size and number of turns, but they don't realize that what works for a soldering iron or a hobby train may not be good for an audio product. Charts tend to result in the most economical product, operating with high induction results in a kind of self-regulating xfmrs, but they are noisy, run hot and radiate hum.
Typically I asked my winders to size the xfmr about 30% larger than what the chart said; operating at reduced induction solved a lot of issues.

 

[Whoops]

What are the voltages with R1=R2=9.1k?

Hi Abbey,
was able to measure the voltages with R1=R2=9.1k, like you asked,
here it is (blue values are old voltages with R1 and R2 22K, and Green are the voltages with R1=R2=9.1k)

 

[Whoops]

I believe the problem is the DCR (and maybe leakage inductance) of the xfmr. Can you measure both primary and secondaries DCR?

here it is both primary and secondaries DCR

 

[doqmemory]

Hi Abbey,
was able to measure the voltages with R1=R2=9.1k, like you asked,
here it is (blues values are old voltages with R1 and R2 22K, and Green are the voltages with R1=R2=9.1k)

View attachment 89909

Does it sound different now?

 

[abbey road d enfer]

Hi Abbey,
was able to measure the voltages with R1=R2=9.1k, like you asked,
here it is (blue values are old voltages with R1 and R2 22K, and Green are the voltages with R1=R2=9.1k)

View attachment 89909

Well, it shows that the zeners are just starting to regulate. However, they make sure that , in teh case of leaving the PSU on and disconnecting teh mic, the output voltage would not exceed the specified 212V.
If noise was an issue. the zener string could be moved across C5, but it would need to be adjusred to 219-220V by adding a 7.5 or 8.2V zzner.

 

[Whoops]

Well, it shows that the zeners are just starting to regulate. However, they make sure that , in teh case of leaving the PSU on and disconnecting teh mic, the output voltage would not exceed the specified 212V.
If noise was an issue. the zener string could be moved across C5, but it would need to be adjusred to 219-220V by adding a 7.5 or 8.2V zzner.

I can easily move the Zener String across C5 and add a 7.5 or 8.2V zener. No problems with that.

What do you think about the DC resistance of the transformer?
I'm still trying to figure out why the transformer seems to be saturating,
B+ consumption is only 1mA and the Heater is only 200mA, the transformer core size is same size or bigger than other 9V transformers I have rated for 1 to 1.5 Amps. So it's still a mystery to me.

Thanks

 

[abbey road d enfer]

What do you think about the DC resistance of the transformer?

It's clear that it's the primary cause for the voltage drop. Probably npot saturation.

 

[Whoops]

It's clear that it's the primary cause for the voltage drop. Probably npot saturation.

Could you please explain why did you reach that conclusion?

What do you think can be done to solve it?

Thanks