Removing U67 Filter Circuit

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
micaddict said:
In76d wrote about the E80F. If it fits on the IOaudio PCB, too, you could give that a try. From what I've heard it's a little more open sounding, but a little leaner on the low mids. The proof is in the pudding

It should fit :)

"a little leaner on the low mids" could be hearing impression affected by little bit lower output.
In G7 which i made, in triode connection, it draw twice more current than any EF86/EF806.
I raised plate voltage by higher value of cathode resistor.
Also i used larger cathode capacitor, finally output capacitor was decreased from 3.3uF to 0.5uF :)
There's no deficiency in the low end or mids.
Clean, noiseless, sweet sounding :)
Here's different circuit but any improvements can be easily made .

 
"a little leaner on the low mids" could be hearing impression affected by little bit lower output.

Yeah, very true. And exactly that is what some audio salesmen take advantage of.
But I always match volume of course.

And thanks for the tech info. That can always come in handy.

Hey, I ordered a matched pair Philips E80F today.
Those puppies will be coming back to where they were born.
That is The Netherlands.  :)
 
micaddict said:
First off, it's your mic, so it's your call, obviously.
Also, if you're able to undo it, you'll likely be able to redo it.

I would strongly advise against it, though. After the treatment, your U67 will no longer be a U67. It will basically be like the countless modern mics with seriously hyped tops. Well, with better parts than most. But nothing special really.
If that's what you're after, you could also get or build such an animal from scratch.

The U67 circuit is a very sophisticated design. Lots of brainpower went in it. And there's nothing like it.
And of course the K67 capsule was designed with that intricate high frequency filtering in mind.  The highs in the K67 are pre-emphasized, so without de-emphasizing the capslue will be overly bright.
A healthy U67  should not sound dark BTW. The original plots are flat, with a very slight rise even, depending on polar pattern. They have been called dark, but that would be as compared to modern LDCs, many of which are not flat. Also, quite a few, if not most of the originals are tired to some extent and should be sent to an expert. They will sound fresher upon return!

Of course you could install a stepped filter. or you could use an equalizer to tame the highs after the filtering circuit has been removed. And yes, often filtering is better than amplifying.
But there are exceptions to that rule and the U67 is one of them. The U67 takes lifting of the highs like no other microphone. That is, if you have a good equalizer.
And with the circuit intact it acts like a(-n active) de-esser (as if there's a compressor in the side chain). Or should I say a de-eshjzer? There's undistorted sibilance and there's distorted sibilance. I'm talking about the latter. Again, here's a paradox. The U67 has lots of negative feedback and often that would mean a less clean/fast/pure path. But it was put there deliberately and the whole of the circuit (with nothing left out) fights distorted sibilance and similar artifacts that can occur in condenser mics. And consequently you can lift the highs more than you would expect, without getting spikes and filth.

This is my miserable way of explaining, mind. Others, like top recording engineers who have worked with U67 when both the mics and the men were in their prime, have done a better job explaining it. Still, it remains an intricate design, made by true masters of the craft (who were given almost unlimited financial means for designing), so it's not so easy to grasp.

Apart from the filtering system as discussed above, there's also a broadcast filter. You have the DU67 and Dany's version is true to the original, so the broadcast filter is there, too. Another U67 cloner here decided to leave it out from the start, which makes his not true to the original in that regard. But that's a lot less radical.
So IF you decide to go ahead and take filters out, I would start with the broadcast filter.

But again, it's your call.



Henk

Resurrecting this topic from a few months ago, I was busy dealing with some family stuff that kept me from working on music for a short while, thank every one for their input!

Henk, I've done some research and it appears that removing the broadcast filters is one of the most common mods that are done to most 67's (as you suggested).  what components do I need to remove to get rid of the broadcast filters?
 
C17 - removing completely will mess level a little, i would replace it with small value cap  - 47pF for example if you want no  LPF. Neumann suggested values are 80-160pF sou you can try few values and see how different it can sound.
If am not wrong C2 and C3 will affect low end response, but i never messed any U67 circuit so i can be wrong here.
Anyway you can try replace C2 with 22nF-68nF.
Still suggesting to try E80F (not related with EF80) ;)
 
Hey, I'd almost forgotten this one.  :)

I think it's best to leave the details now to others more qualified than I am.
There's Dany (poctop) who  leaves the broadcast/RF filters in to stay truest to the original, but obviously knows exactly how to remove them.
There's Max (ioaudio) who chooses to take or leave them out from the get go.
And our new fellow member Klaus Heyne has a reputation for "opening up" U67s without butchering them. He might chime in, too.
Just to mention a couple.



 
Ioaudio circuit doesn't remove the filters, maybe I misunderstand your meaning. 

Max posted this at GS as a mod for his package, use the option to return to full thread.

https://www.gearslutz.com/board/11999894-post475.html

This shunts point C2/C3/R3/R4/R5 to ground, and is likely very similar to the standard C17/S2 modifications.  If C3 and C5 are next to each other as in an original U67, you can do a quick test by shorting the correct sides of those together.  I've done it on an internal jumper at that position to have both available, and have the HPF available when in standard mode.  The HPF won't do anything useful in modified mode. 
 
Ioaudio circuit doesn't remove the filters. 

I don't mean the de-emphasis filter, in case anyone might assume that.


Max wrote something about it, but i can't find it all right now.
This bit I did find:
i leave out the feed-through caps on the bottom, since there's enough filtering going on on the rest of the circuit.
 
I prepared myself to discuss a little bit :D

Yes, RF caps (C12, C13) can be removed, in case of any problem with RF (shouldn't be) there's always option to put it back.

Broadcast filter, i believe, is  30Hz roll off, which can be omitted removing S2 jumper.

Back to original question:

tesco_1 said:
I have a DU67 with a genuine Neumann Capsule w Nos/ valvo/telefunken tubes, and was interested in removing the high frequency filter circuit to open the top end of the mic and increase clairty.  Or even making a switch to be able to change between a brighter more open modded 67 sound and the stock 67 sound.  Can anyone tell me what components would I need to remove to remove the HF feedback circuit?

Additional switch could be best soulution.
You could keep original response and another, modeled for your needs.

I would use three position switch DP3T and make three configurations:

- original
- more open highs
- more open highs and removing low end roll off

Or two configurations with DPDT:

- original
- more open highs and removing low end roll off

This way you can have microphone for different purposes - i would call it "classic" and "modern" :D

C17A and C17B can have different ratio.
Best would be make test and empirically choose values to get total "160pF".
 

Attachments

  • U67 switched response mod .jpg
    U67 switched response mod .jpg
    242.7 KB
You want less pfd, not more, for extended highs.  The classic mod for more highs is complete removal of C17. 
 
emrr said:
You want less pfd, not more, for extended highs.  The classic mod for more highs is complete removal of C17.

So you have less pF. For start of HF boosting is a half value of nominal 160pF, and as i wrote best would be make a test to find most usefull values.
With half value you will notice much hf boost. Same with u87.  It can be different ratio like 47pF C17A and 110pF C17B
Removing C17 completely will remove a part of feedback, jumpering will put DC in the path. This way is much more simple and it change response without changing output level. Still you can use original response eq, so it's in my opinion much better than removing feedback completely ;) For the microphones without deemphasis feedback i know a lot better circuits :p U67 main advantage is exactly that part of the circuit! Also! Removing RF caps, still would be good to keep some HF filtering!
Switch position from below to top:
- original
- more open highs
- more open highs and removing low end roll off
 
OK, I see where you are coming from.  Most modern DIY builds appear to be 100pfd.  I don't hear much difference taking it down to about 70pfd, but will take some more comparative measurements soon. 
 
I didn't know :D
Am not interested in cloning U67, i took original schematic :)
Good that you mentioned that ;)
I only repaired U67 but didn't "tweak" it response. In U87 the small changes in values gives audible differnces.
Between 100pF and 70pF difference shouldn't be that much. Anyway 100pF shouldn't affect  much HF response.
Sticking to this value i would go for double 47pF.
Anyway if  tesco_1 would  use additional switch option i think it could be good too keep 160pF value - then the difference between each setup should be really audible - it's good option for multiple types of voices or different types sources, isn't?
 
Yes, any modification may be useful for some setting, and the user might best tune it to their mic and needs. 

I assume the variance in the value has to do with tuning for the particular transformer/amplifier combination, since the manual gives an amplifier test with acceptable results, and the variance is not meant to address the capsule.  But I don't know, and I don't know if there is a trend in C17 value found in the field on real U67's. 
 
30Hz LPF is a rumble filter rather than a "broadcast filter". And RF may be a better term for the latter.
And  wasn't 15K LPF done to avoid the FM pilot tone?

I personally would never mess with the de-emphasis/negative feedback circuit, because that's where the magic of this mic is IMO.
 
micaddict said:
30Hz LPF is a rumble filter rather than a "broadcast filter". And RF may be a better term for the latter.
And  wasn't 15K LPF done to avoid the FM pilot tone?

I personally would never mess with the de-emphasis/negative feedback circuit, because that's where the magic of this mic is IMO.

Hey Henk :)
As you can see i did the simple mod to keep both - original deemphasis values and something which could fit the needs from the first topic.
With the "broadcast filters" in U67 first time i read here.
If it's not permanent HPF (not LPF ;) ) which you called rumble filter, i don't see any other.
RF is RF, it shouldn't affect 15kHz, rather much higher since caps are 2nF. 
15kHz is area where deemphasis eq work, so if Neumann designed it with that assumption, it should work with 160pF, and maybe that's why is popular these days to use 100pF as emrr mentioned.
 
Once again thank you all for the swift replies!  Tons of info to look into.  So a few days ago I tried the quick and dirty C17 removal to see what would happen, and I put it right back! 

The untamed high frequency boost was for lack of a better term 'ugly', and unbalanced.  Although the high frequency boost was not incredibly bright (maybe due to the IO transformer?) it was notably different from the stock mic, and had a non linear distortion thing going on up there.  I definitively may experiment with other C17 values via a switch as suggested.

Also I am aware of the S2 low cut jumper, and it's relatively irrelevant to me as I mostly record vocals, and it's already switchable.

Now to get some things straight:  "RF filters" is the correct term for what some call the "broadcast filters" (implemented in compliance with German radio specs)? If I understand correctly removing this filter should eliminate the high-end roll off (I assume around 14k)? And again what components make up these RF filters?

Ultimately I just want to get a little more top end out of the mic with respect to its general design.


 
I recently put a metallized paper output cap in mine and it gave it a more "musical" high end. It now doesn't sound as dark. I still have a lot of testing to do before I decide whether I'm keeping it that way or not. It sounded better on a few vocal tests I did but I worry that it might make that mic less versatile.

Might be worth a shot.
 
emrr said:
Ioaudio circuit doesn't remove the filters, maybe I misunderstand your meaning. 

Max posted this at GS as a mod for his package, use the option to return to full thread.

https://www.gearslutz.com/board/11999894-post475.html

This shunts point C2/C3/R3/R4/R5 to ground, and is likely very similar to the standard C17/S2 modifications.  If C3 and C5 are next to each other as in an original U67, you can do a quick test by shorting the correct sides of those together.  I've done it on an internal jumper at that position to have both available, and have the HPF available when in standard mode.  The HPF won't do anything useful in modified mode.

I checked out the thread on GS, thanks.  I still relatively a novice when it comes to electronics, I've built some pre amps, and mics, but I get by mostly by just following instructions and attention to detail.

Can anyone tell me how would I perform this linear mod with Danny's pcb?


 
Ultimately I just want to get a little more top end out of the mic with respect to its general design.

Just remember many oldtime users have found that no condenser mic takes a high frequency  EQ lift like a U67. (Post microphone that is.)
So basically, tempering with any part of the de-emphasis/negative feedback circuit is like throwing out the baby with the bath water.
Or worse.

RF filters are another matter, of course, as is replacing a cap for another type (not value).
 
Back
Top