Warm Audio WA-67 - Teardown

[Whoops]

What makes you think the GAP 47 uses a K67-design? Could be a K47 in 35mm, no?

This is a Warm Audio WA67 thread, can you please not Hijack this thread?
Please

You can ask that in private to Kingkorg or do a new thread on that subject that is not at all related to the Warm Audio WA67

 

[Whoops]

This may of some interest in selecting output cap size:
Coupling Capacitor Calculator by V-Caps
Maybe they are compensating for the transformer.

Thank you so much for the calculator mate,
although I'm only a Sound Engineer that likes to DIY gear as an hobbie, I'm not trained in electronics,
so I don't really know how to use that

Thanks

 

[pinchemotherloaf]

Thank you so much for the calculator mate,
although I'm only a Sound Engineer that likes to DIY gear as an hobbie, I'm not trained in electronics,
so I don't really know how to use that

Thanks

Load Input Impedance (Ohms):	10000
-3db Frequency (Hz):	20
Capacitor Value (µF):	0.80
Optimal low frequency response (Hz):	200.00

Here's an example by entering the transformer (network) impedance of 10000 ohms and roll-off frequency of -3 dB at 20Hz would require a cap near .8uF
Changing the transformer impedance to 2000 ohms requires a cap around 3.98uF for the same roll-off frequency at 20Hz. Maybe the WA mic has a transformer with the lower impedance requiring a larger cap.

Load Input Impedance (Ohms):	2000
-3db Frequency (Hz):	20
Capacitor Value (µF):	3.98
Optimal low frequency response (Hz):	200.00

 

[kingkorg]

Well, the picture is clear enough, it's really a 4.7uF that's installed in the Warm.
As to the reason, I can only speculate. Maybe they did it preventatively, maybe they did it because their xfmr has a lower inductance (but I doubt by a factor 10), maybe they made a mistake, maybe they had only this value in stock...
In order to know more, either someone can pick the designer's brains, or someone measures the circuit's response and the xfmr separately.

Back to the topic.
Most of the times people use larger caps here to get more low end, maybe avoid resonance in audible range. Same for the high value resistors, which can be detrimental in tube circuit. Lundahl made a custom transformer for them, i doubt they missed the spec if they went for a replica.

@Whoops, have you checked the high value resistors? My guess would be they "upgraded" those to 1G value.

At this point only idea i have for this mic is to measure the specs and compare to u67, mod it to match the performance of the original. But as i am sure it works just fine as is, maybe just leave it alone. Or add an internal switch for multiple voicings with diferent feedback cap values. Or even dual output cap.

Warm audio lacks another crucial element that affects the sound big time. The dome under the capsule.

 

[Whoops]

@Whoops, have you checked the high value resistors? My guess would be they "upgraded" those to 1G value.

Hi mate,
I checked all component values when I traced the schematic, the only values changed are the ones that are in "Red Color" in the schematic, all the other ones have exactly the same value as the original Neumann schematic for the U67

 

[kingkorg]

You were able to measure R1 and R8 in circuit with a regular meter? And they measured 400meg?

 

[soliloqueen]

Back to the topic.
Most of the times people use larger caps here to get more low end, maybe avoid resonance in audible range. Same for the high value resistors, which can be detrimental in tube circuit. Lundahl made a custom transformer for them, i doubt they missed the spec if they went for a replica.

@Whoops, have you checked the high value resistors? My guess would be they "upgraded" those to 1G value.

At this point only idea i have for this mic is to measure the specs and compare to u67, mod it to match the performance of the original. But as i am sure it works just fine as is, maybe just leave it alone. Or add an internal switch for multiple voicings with diferent feedback cap values. Or even dual output cap.

Warm audio lacks another crucial element that affects the sound big time. The dome under the capsule.

WA-87 sounds wonk above 6k too. standing wave issues. incorrect head basket size and contents would make sense

 

[Whoops]

You were able to measure R1 and R8 in circuit with a regular meter? And they measured 400meg?

I was able to read the color code on all the resistors...

 

[Whoops]

Warm audio lacks another crucial element that affects the sound big time. The dome under the capsule.

I thought about that before also.
But Warm Audio not doing it, doesn't shock me, but I was quite surprised that even Neumann is not doing that in their U67 Reissue

Left is the U67 Reissue, right is a Vintage U67

 

[Whoops]

I would have no difficulty in 3D printing a dome and install it in the WA67,
might be something to do now that I think about it

 

[soliloqueen]

I thought about that before also.
But Warm Audio not doing it, doesn't shock me, but I was quite surprised that even Neumann is not doing that in their U67 Reissue

Left is the U67 Reissue, right is a Vintage U67

View attachment 88597

I found out about this when I was researching u87 revisions. They decided it was more efficient to break up the standing wave by simply moving the capsule up in the headbasket. It's pretty effective, but it doesn't sound exactly the same as the dome.

 

[Whoops]

At this point only idea i have for this mic is to measure the specs and compare to u67, mod it to match the performance of the original. But as i am sure it works just fine as is, maybe just leave it alone.

Yes, I'm not going to change a lot,
I was just curious about the reason for the 4,7uf cap.

I fixed the capsule mount, changed the way the 1000pf caps where mounted (purely visual), I retouched all solders. I did some tests with some dampening on the body, I have a video to post here showing before and after. And i'm not going to do much more

Probably just checking the PSU because the B+ voltage seems really low, I measured +191V instead of +210V (+- 2%).
I also don't know the reason for it to be set so low, there's a trimmer in the PSU, I will check it that trimmer is a B+ voltage adjustment or not.

Besides that I don't want to change anything else for the moment,
maybe trying some different tubes because I have some stock of NOS EF86, and I'm curious to test if changing the tube really makes a big difference in sound or if it's marginal (for my needs)

 

[Bart_J86]

The u67 is shown as 0db flat line in the graph. Which means it is used as reference. Yellow line is wa67's deviation from u67. The difference, which is +-5db, total of about 10db in difference. In my book they couldn't be more different.

When it comes to your screenshot graphs what you see as common are huge peaks and dips which come from surroundings. Not from the mics. Naturally they look the same as you are recording in the same space. However you are missing the details where they differ and deviate from each other. And they differ a lot!

Do a 1/3 smoothing on those graphs and you will see what i'm talking about. Or use something like curveEQ to show you the difference between the two responses.

You are also using 10+db per division, which is not helping.

Edit:
Nevermind, i did it for you. The CurveEQ image below shows the difference between the two with cardioid strumming example. I hope you get the point.

Hi all, I’m rather new to the diy world and really enjoying the way you analyse this mic and it’s construction. Hopefully I’m not hijacking this thread with this rather basic question, but can someone explain to me how to make such a frequency spectrum analysis graph (which soft- or hardware do you use?), and how do you get reliable results on a mic? Would be very grateful if someone is willing to explain this to me as I would really like to test the diy stuff I build so far

 

[abbey road d enfer]

This may of some interest in selecting output cap size:
Coupling Capacitor Calculator by V-Caps
Maybe they are compensating for the transformer.

This is for a basic CR coupling. We are discussing CL coupling.

 

[Khron]

Hi all, I’m rather new to the diy world and really enjoying the way you analyse this mic and it’s construction. Hopefully I’m not hijacking this thread with this rather basic question, but can someone explain to me how to make such a frequency spectrum analysis graph (which soft- or hardware do you use?), and how do you get reliable results on a mic? Would be very grateful if someone is willing to explain this to me as I would really like to test the diy stuff I build so far

The easiest and cheapest way is to use whatever audio interface you have; good-enough USB ones can be had for as little as $50 used (or even less, if you're patient with your hunt).

Software-wise, RoomEQ Wizard (aka REW) is free and VERY capable. In an ideal world, you'd want to have a (calibrated) "reference" microphone, which is assumed to have a flat response. Using that as a "known value", you can then compare whatever other mics with that "yardstick". But if relative measurements between mics are enough for you, you don't necessarily need the calibrated reference.

 

[Bart_J86]

The easiest and cheapest way is to use whatever audio interface you have; good-enough USB ones can be had for as little as $50 used (or even less, if you're patient with your hunt).

Software-wise, RoomEQ Wizard (aka REW) is free and VERY capable. In an ideal world, you'd want to have a (calibrated) "reference" microphone, which is assumed to have a flat response. Using that as a "known value", you can then compare whatever other mics with that "yardstick". But if relative measurements between mics are enough for you, you don't necessarily need the calibrated reference.

Thanks a lot! I’ll definately try the REW software which looks promising and just what I was looking for. I’ve got a SoundID Reference measurement mic (I hope I can download the corresponding calibration file) and a focusrite clarett 8pre, so i’ll run my diy projects into the REW software for some measurements

 

[abbey road d enfer]

Thanks a lot! I’ll definately try the REW software which looks promising and just what I was looking for. I’ve got a SoundID Reference measurement mic (I hope I can download the corresponding calibration file) and a focusrite clarett 8pre, so i’ll run my diy projects into the REW software for some measurements

Beware that REW, although excellent, requires some knowledge of measurement theory. You need to know how the number of FFT points interferes with the LF response. Actually, as the frequency decreases, measurements are less and less reliable. Reading the REW help is not to be neglected.
OTOH, HF measurements are very sensitive to mic placement; in particular, one must clearly identify where the diaphragm is located and make sure the position is the same in each measurement. Placing two mics close to each other does not work correctly.

 

[amplexus]

What i hate the most about these is that way too many companies claimed to sell "premium" or "selected" or "tuned" versions for more money. When in reality there is pretty much nothing one can do to make them perform as a proper k67 should. First of all overall diameter is wrong in both models. Some of these companies are Telefunken, Peluso, AA... And way too many are sold in some very expensive models. SE RNT for example. Yikes that one sends chills down my spine

Sorry, but I have to take issue with this claim wrt to Peluso and AA using generic cheap off the shelf capsules. I happen to know a bit about their supply chain, and they are absolutely not doing this. Certainly all of Peluso’s capsules are being skinned and tuned in the US, mostly by John himself, with the metalwork being done offshore to custom spec.

 

[Whoops]

Besides the 4,7uf output capacitor,
I'm also quite curious as to why did Warm Audio set the B+ voltage for the WA67 at 190V instead of 210v.
It seems like a big difference to me.

Do you guys have ideas on what might be the reasons for it? why the 190V?

 

[Khron]

Wouldn't be 10000% shocked if mains variations / tolerances would be involved, especially since the mains-to-secondary ratio is 1:1 or greater. Did you measure your mains voltage at the time?