Eckmiller W86 - refurb and makeup amplifier options

[telewulf]

The W66 is passive so its input and output impedance is the same (200 Ohms)

Ooops! Wrong again!

 

[thomasdf]

Wow. Thanks for sharing so much knowledge !

You are making me rethink my whole approach here. Should I go IC instead of transformers and Class A … I don’t know how I can achieve 20 ohms and 200 ohms without transformers. I’ll look into it.

BTW, your EQs are 86a, which are inductor based. Mine are 86, they are RC.
You have a lovely setup !! I am jealous haha

 

[telewulf]

You could evolve it into a two step program.
Step 1: Try the W86 in a quick but correct op-amp environmet and see if the Eckmillers are working as they should.
Step 2: Build a transformer coupled high brow amp later.

how I can achieve 20 ohms and 200 ohms without transformers

Well it does not have to be precisely 20 Ohms, just the lower the better.

I may be completely wrong, but it reminds me of a power amp that is driving an speaker. Every speaker has its resonant frequency an when you hit it you get coloration but you will have the physical mass of the cone performing unwanted oscillation as well, like a pendulum that resonates at its own resonant frequency. So what to do about that?
Well, your amp has to be powerful enough to force the cone´s mass to perform exactly the movements it ought to an not to resonate as it wants to.
Same thing with your W86: It has its own resonances and you have to supress them by driving it with a powerful source and as well by making it drive the correct load.

Here´s what I would do as "Strep 1":
The correct load is easy enough to achieve: Every mic-input on every mixer has an input impedance of 2 kOhms and thus provides you with the correct load automaticly ... and it´s symmetrical as well ... and it gives you your 34 dB too - bingo!
As for the driver: To provide 1,55 V with an source impedance of 20 Ohms you need 0,12 Watts, so with two TDA2822 on opposing phases you ought to be on the save side. You might try these:
https://www.ebay.de/sch/i.html?_from=R40&_trksid=p2332490.m570.l1313&_nkw=TDA2822M&_sacat=0
... or something similar
That shouldn´t take more than one evenings work should it? And it would give you good results on how well or not your Eckmillers are.

best wishes from Bremen,
Wulf

 

[telewulf]

your EQs are 86a, which are inductor based. Mine are 86, they are RC

Yes, and i NEVER knew that!
I only wondered why they looked so very differently although bearing almost identical names. So the tread you started has been very informative for me es well - thanx a buch for that!

 

[telewulf]

Oh, and by the way: Only one of my Eckmillers is a W86a, the other two very similar looking units are HV55. HV stands for "HörspielVerzerrer" wich would translate as "Distorter for radio plays". Just like the HS10 and TS10 they would just make a hi- and lo-cut, but at rather insane frequencies. For instance with the HV55 the lo-cut could be raised up to 6,4 kHz and the hi-cut could be lowered to 200 Hz - nothing to make your music sound nice, but to make it unrecognizable!

The HV55 even had a brother, the HV53 wich again looks just like your W86 but with the insane frequencies. And then there was a third "Hörspielverzerrer", only it wasn´t named "HV" and wasn´t made by Berlin based Hans Eckmiller but by Hamburg based Hugo Maihak. It was the W49 - a huuuge machine that was only produced in about 100 specimens for the NWDR wich in 1954 split into the NDR in Hamburg and the WDR in Köln, where they installed the magnificent Studio für elektronische Klanggestaltung" - Studio for electronical sound design at the WDR, a kind of radiophonic workshop, where a jung German composer named Karl Heinz Stockhausen fell in love with it - so much so, that he even wrote its handling into his sheet music.
You find a photo of an entranced young Karl Heinz attached, the device nearest to the photographer is one W49 and anothere one on the other side of the special built mixer.
BUT: the Eckmiller HV55 has one advantage over the much grander W49: You can choose the frequencies freely as you wish, whereas with the W49, you had two levers in the same groove and thus you could never put the lo-cut higher than the hi-cut.

best from Bremen,
Wulf

 

[Dark1]

[...TS10] So I bought on unit of each because I knew even then that they were built to be used only in conjunction with one another. They as well are completely passive but since they do only cut, they do not require a 34 dB boost. So I racked them and fed them into a Maihak W66. The resonances were so clearly audible, I did not even have to make any measurements. I was so disapointed and sold them emidiately.

Back then Eckmiller also supplied transformers (CT10/11) in case they were not used together, to present a required impedance. But I have never seen one, sadly; using them strapped together was always very convenient, and cost-saving solution I guess.

Gosh, am I lucky to have stumbled onto this thread! Great infos (thank you thomasdf!) and @ Rock: What a wonderful leaflet!!!

When I started work at NDR in Hamburg there was still one small studio equiped with Siemens V72, TAB V76, Maihak W87 and Eckmiller EQs. When it went out of comission I was lucky enough to be able to buy the console and it´s serving me to this very day - longer than it has served German Radio by now and without one day of failiure.
(I could not blame you if you accused me for boasting, so: Sorry in advance!)

Amazing! Please tell me you also saved documentation of the console (...and intend to scan it someday)? It is very rare to see such a console not cut up to pieces and sold for modules, and it would be even more rare to see original docs (console wiring diagrams, test reports etc.). In any case, very admirable to save the console and keep it in use!

 

[telewulf]

Please tell me you also saved documentation of the console

Yes I did
And luckily so, because at the NDR the thing had been changed so many times and I would have been at a loss to know left from right.

(...and intend to scan it someday)?

Why not? But it is many big sheets of paper because they painted a new block diagramm with each modification ... in ink ... by hand. Ooo, all the work they put into it! So of course I wouldn´t want to spoil their legacy and continued to record every alteration I made.
For instance when I worked at it in 1991, it had lost the V76 it once contained and I wouldn´t even have noticed if it hadn´t been for the old block diagramm from 1957.

Well, and when it was to be torn out in 1993 one builder told me, there would be a crane coming the next day to heave it from the second floor and drop it into the container ... but if he accidently found that there would be a trailer being parked right next to the container, he might as well let it down gently two meters next to the container.

And when I moved from Hamburg to Bremen a couple of years later, a friend of mine, whom I had asked if he could help me, replied: "Do you still have that massive thing? Size of a washing machine but twice the weight?"
"Yes" I said "I still own it."
"Well then I will not help you." he replied.

 

[telewulf]

Back then Eckmiller also supplied transformers (CT10/11) in case they were not used together, to present a required impedance.

Aaah! I didn´t know that eighther! I always wondered how there could only be a TS10 in the side rack to an old console of "Deutsche Phonogramm" that a friend of mine owns, without the adjacent HS10 - well, that explains it, thanky you!
That console - by the way - does contain the V72´s little brother: the V71.

 

[Dark1]

Yes I did
And luckily so, because at the NDR the thing had been changed so many times and I would have been at a loss to know left from right.

Why not? But it is many big sheets of paper because they painted a new block diagramm with each modification ... in ink ... by hand. Ooo, all the work they put into it! So of course I wouldn´t want to spoil their legacy and continued to record every alteration I made. [...]

That is absolutely amazing! I would gladly help with scanning that if you can't or don't have time to do so (I have a small scanner but can stitch the images in Photoshop together, no problem with that). Or I can clean and stitch the scanned images, whatever suits you (PM me). I can also understand you won't want to bother with that at this time, it is some work, but please don't throw it away... This is a treasure trove of information esp. on gainstaging and impedances, years of manufacture etc.

Aaah! I didn´t know that eighther! I always wondered how there could only be a TS10 in the side rack to an old console of "Deutsche Phonogramm" that a friend of mine owns, without the adjacent HS10 - well, that explains it, thanky you!
That console - by the way - does contain the V72´s little brother: the V71.

Malotki V71/V71a are very rare, and information on them is also pretty rare as well esp. information on different revisions (V71 w/ 19/34dB Switch, V71 w/out Switch, V71s, V71u)...

 

[Dark1]

[...] I need a 34dB makeup amp after them (or before, from what I read it can bring a nice color). The thing is they are passive, balanced and need to be fed and drive symetrical signals. I am wondering what path I could follow.
Back then it was designed to be used with a V72 but that's out of my budget and I would like the amp to fit in a 1U rack unit.
I have 2520 makeup boards that would need both an input and output transformer
I thought of giving the Neve BA283 a try, too.[...]

Siemens Sitral V272 would be a good match, and still affordable?

 

[thomasdf]

May I draw your attention to page 8 of the lovely leaflet? They give specific information about the source and load impedances and they are tremendously importand if you want your Eckmillers to work properly!
Since they are completely passive, they interact heavily with the source by that they are driven and with the load that they themselves drive.

@ thomasdf:
And for that reason it is of the utmost importance, that you use the W86 in an environment where they are driven by a source impedance of ca 20 Ohms or less (a V72 does precisely that) AND where the load is ca 2 kOhms (V72 again) but not more than the stated 4 kOhms. The trouble with buffer amps with an input transformer might be, that it might lean too much towards being an inductive load, which will alter the frequency curves of your W86. You´d be much more on the save side if you used an op-amp-buffer with a high input impedance and strapped 4,7 kOhms across the input.

All these old German modules were built to interact with one another and the passive Eckmiller-EQs even more so!

As I have some spare time this summer I am starting to rethink about this project and your input is much welcome, and enlightening. If I am understanding correctly I need to feed the EQ a precise load, and they need to see a precise load at their output, right ?
That would be under 800 ohms input impedance, with a 4K load / strap and 200 ohms output impedance with a 25R load ?

So that would mean : IC or solid state input buffer - EQ module - microphone style preamplifier with 34dB of makeup gain.

 

[thomasdf]

Siemens Sitral V272 would be a good match, and still affordable?

A little on the expensive side for my wallet, but classy for sure !

 

[telewulf]

Hello Thomas,

good to hear you are working on the old Eckmillers again!

If I am understanding correctly I need to feed the EQ a precise load, and they need to see a precise load at their output, right ?

Yes, that´s what i meant. Easiest way would be:
- take a head phone amp
- feed the left channel with a signal generator
- feed the right channel with the same signal but with a phase of 180°.
- feed the the outputs pin and ring to the W86, leavin out the sleave.
- feed the output of the W86 to a mic input that has an input impedance of about 2kOhms.
- have fun with the measurements

If you want to have even more fun, you might try the set up in the picture. You than can vary the input and output impedances and see how it might affect the frequency response, especially in respect to any resonance-humps in the plot, that might or might not be reduced when you reduce the impedances.

hope you have fun!

Best wishes from Bremen,
Wulf

 

[thomasdf]

I am not sure I understand the 2KOhms impedance preamp point, from what I understand from the manual it calls for a 200 ohms preamp ?

I did some testing with a headphone amp and a little Mackie desk I had around, the EQs work as expected but the tone was not particularly pleasant or lush, and overall the was a slight tilted towards to high freqs, with this less than ideal setup.

 

[telewulf]

Well, then let me explain the basics of impedance matching and forgive me, if that is information you already have, but it´s hard for me to guess, what or what not you already know. So if you are familiar with everything concerning impedance matching, you can skip what stands between the lines
______________________________________________________

There are three different kinds of impedance matching:

current matching:
You want to get the biggest possible current from the source to the destination. So the input resistance has to be as low as possible and by all meany considerably lower than the souce impedance.
=> Rout > 10 x Rin
In audio apliances we only find that in summing amps and because of the low input impedance it is referred to as "0-Ohm-technique"

voltage matching:
You want to get the biggest possible voltage from the source to the destination. Think of Rout and Rin as two resistors in series. If you want the the voltage drop across Rin to be as big as possible, Rin itself has to be as big as possible
=> Rout < 10 x Rin
This is the kind of matching that we usualy encounter in audio apliances and also in our case.

power matching:
You want the tansfered power to be as big as possible. Power is the product of current and voltage and it turns out that the transfered power is at its maximum when Rout and Rin are equal.
=> Rout = Rin
We only encounter this kind of matching when transfering power from a power amp to a loudspeaker.
______________________________________________________

Now here is a translation of the last page of the W86-manual that Rock Soderstrom kindly provided:

Circuit: Symmetrical design in H-shape, according to the principle diagram above, Figure 4, using RC elements.

Input resistance equal or greater than 800 Ohm in a frequency range from 0 to 15 kHz and at any switch position, with an output measurement termination of 4 kOhm.

Output resistance = 200 Ohm +/- 20%, in a frequency range from 0 to 15 kHz and at any switch position, with an input measurement termination of 25 Ohm.

Basic attenuation = 34 db +/- 1 db, in a frequency range from 0 to 15 kHz and position 0 db the switches "60 Hz" and "10 kHz".

Frequency deviation from the attenuation curve shown in Figure 5 in the individual positions of the switches "60 Hz" and "10kHz" is lower than or equal to +/- 2 db, in a frequency range from 0 to 15 kHz and position 0 db the switches "60 Hz" and"10 kHz."

So if the W86´s input impedance is 800 Ohm and voltage matching aplies, then it follows:
Rout of the driving amp must be 80 Ohms or lower.

And if the W86´s output impedance is 200 Ohm, then it follows:
Rin of the follow up amp must be 2000 Ohms or greater.

And because the W86 is passive ant thus very-very-very reliant on correct matching conditiones, they even give information on the measuring environment they themselves used:
Rout of the driving amp was 25 Ohms
Rin of the follow up amp was 4700 Ohms

Did this information straighten things out a bit?

Best wishes from Bremen,
Wulf

 

[thomasdf]

It does Wulf, thanks a lot ! Allowed me to clarify some stuff I did know about and for some reason didn't really explain. I guess I've also missed some info from the german brochure, obviously. I am back at my drawing board. I'd like the design around the EQ to be as elegant as possible, and I like the idea of it being all balanced, too ... I'll come back with some schematics soon !

Thanks again

Thomas

 

[thomasdf]

Here is what I have so far... I assembled stuff found in Douglas Self's book and online ... so maybe it's totally crap ? At this point I am not very sure.

My goal is transformer > unity gain buffer > W86 > makeup gain > transformer. I like the idea of having a completely balanced circuit, but I am not sure how to set the correct impedances and load... or use the correct value coupling caps. Maybe I don't even need C2 / C1

The output stage is taken from a D Self design where the primary of the output transformer is grounded on one side ... so here again, not sure I can do that, in terms of current etc.

 

[telewulf]

Hello Thomas,

sorry for my delayed answer, I hadn´t followed the thread for the last days.

Not crap at all! But it seems to me, you are mixing up symmetrical and asymmetrical signals. The source-transformer gives you a ground-free-symmetrical signal whereas each of the following 5534´s is connected in an asymmetrical fashion. Similar thing at the output.

Now I could expound in two diffent ways:

a) if you are familiar with the intricacees of symmetrical and asymmetrical signals in principle, I could just make some suggestions for alterations or

b) I could first write another long winded reply about the principles of the thing and then make suggestions.

Which way would you prefer?

Best wishes from Bremen,
Wulf

 

[thomasdf]

Haha Wulf ! You are the best ! I should buy you a beer sometime.

I've always assumed a balanced signal was 2 times the same signal, with the polarity opposed and thus could be treated like two unbalanced signals running in parallel ... but now that you mention it, I remember that the Langevin AM16 has some sort of "side to side" connection / relation ...
I realise now that such a design cannot be random parallel floating circuits ...

As said earlier, I like the idea of having a balanced circuit throughout the whole signal path, with the least amount possible of capacitors... Don't ask me why, but I think it's more "classy" and is a nice hommage to the W86 design.

I went for 5534's because I can found some early ones easily, they sound great and are sturdy but I am interested in any other option / idea : Class A, transistor and such...