Rohde und Schwarz Begrenzungsverstärker ABR / U23

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Very excited about this project. I checked the files, already sounding amazing. Well done, very impressive.
 
Dear Jans,

thank you. Since I recorded the files I learnt a lot about this device and took it some steps further. I am still waiting for the new parts to arrive. I already designed a new revision of the main boards and will build them until mid of this year. There are still some details to be figured out but its going to be a really nice compressor.

Best,
falk
 
Good Morning everybody. I am working on revision boards of the units and still try to cure everything that I notice. I invented that small circuit to for the rectified side chain signal. Its super simple and uses a voltage divider R3 / R4 to change the ratio. It works super and as intend.

BUT - in the middle postions when R3 ≈ R4 a 50Hz noise in the audio signal increases by 6dB. (side note: its always R3 + R4 = 1Meg)
The Values of R1, R2, R3, and R4 are all changed by switches.

Does anyone has an idea on this?

Thank you,
Falk


IMG_9952.JPG
 
High impedance node will easily pick up hum and buzz. Try moving the "to grid" wire a bit or routing it differently, it might help or make things worse.
 
Agree, at center of pot you get the highest source impedance. Perhaps try shielded cable here, shield grounded to ground at the divider?

Ideally, any hum introduced here should end up cancelling when added evenly to the two differential grids. Perhaps your balance adjustment is a bit off?

And again, agree: Ratio is just gain/attenuation of control-dc AFTER rectification, threshold is gain/attenuation of signal-ac BEFORE rectifier

/Jakob E.
 
Agreed too. -6dB spot is the highest Z so, shielded cable, playing with the routing of that line, and making sure the audio balance is good.

6AL5 is nearby and AC heaters or? On a PCB or twisted pair? Careful routing of that 6V3 line too if so.
 
Thank you very much for your feedback. I will try and see if I can manage to get it better. The cables to the ratio switch are not shielded so this is going to be the first thing I will check. The dc control voltage is fed through the input tx secondaries center tap. The secondary does not measure ideally symmetrical. In the revision boards I could integrate an option to feed the voltage through resistors instead of the transformer which would improve symmetry. On the other hand it would mean to lose the RL-lowpass.

The 6AL5 is not really near to the cables. All heaters are 6.3V AC with twisted pair. I just realized that there is another 12V AC line on the same multi pin connector that goes to the grids. I ll check if the hum disappears if I disconnect the 12V there. This 12V line feeds the grids in balancing mode.

Other than this there will be 15K:600r output tx´s coming end of the month. This will actually be one of the last modifications.

If this project is going to become DIY I am sure you would appreciate to have a regulated PSU instead of the old fashioned tube PSU right? Is there any regulated PSU draft that could provide 250V and 63V?

Noise is pretty even at -90dB(FS) only 50Hz peaks at around -85dB(FS) when ratio is in the middle position. I use the units almost every day since half a year and they do a great job.
 
Good Morning everybody. I am working on revision boards of the units and still try to cure everything that I notice. I invented that small circuit to for the rectified side chain signal. Its super simple and uses a voltage divider R3 / R4 to change the ratio. It works super and as intend.

BUT - in the middle postions when R3 ≈ R4 a 50Hz noise in the audio signal increases by 6dB. (side note: its always R3 + R4 = 1Meg)
The Values of R1, R2, R3, and R4 are all changed by switches.

Does anyone has an idea on this?

Thank you,
Falk


View attachment 80697
If you look at it as a (high impedance) conductor loop then it´s clear that it will pick up magnetic field. Check your layout so that the area inside the loop is as small as possible. That will lower the induced voltage according to the law of inductivity Uind.=dPhi/dt= d(B*A)/dt
 
Good evening.
Its been a long time with no news because parts took ages to arrive. I have been waiting for 5:1 XSM transformers to replace the 1:1 WSM on the output. But there has been no improvement so I ll stick with the 1:1 WSM. In fact the 1:1 is simply better.

In this mood of disappointment I tried the 5:1 XSM in the position of the 4:1 WSM in the side chain which cured the high pass behavior of the side chain.
The unit seemed to have a 90Hz high pass built in. That has not been a flaw to me because the compression was just right. But if desired the unit does compress equally at lower frequencies in the side chain too from now on.

I still have some distortion issues with electronic kick drums. But I am on it and will figure it out sooner or later. I still need to change the input impedance from 4k to 10k. Is there anything wrong with rising the impedance to lets say 20k?

Also I read the thread about the PM 670 which Michael Kingston improved. My attention was mostly attracted by the PSU. I am pretty sure I will make a regulated PSU with a dc filament supply. I put together parts from Mr. Kingstons schematic and some other schematic I found. It culminates to the schematic attached. Any ideas on this on? I have no idea how a good PSU should be but the Spice simulation looks as if should work :)

It is great fun to have that super heavy choke and regulator tube on the PSU but also its just expensive and a regulated PSU will improve the noise level.
Has anyone an idea for a power transformer that could deliver 120mA @ 225 - 250V and 4A @ 12V ?

I ll keep you posted.
Thanks in advance,
Falk
 

Attachments

  • Bildschirmfoto 2021-07-01 um 23.45.55.png
    Bildschirmfoto 2021-07-01 um 23.45.55.png
    314.3 KB · Views: 86
Last edited:
4007 diodes are only rated 1A, you need much beefier - perhaps 5400-series?

I don't really like the overkill of the 3x 6800uF primary reservoir capacitors C1-3-4 - these drag unnecessary current spikes from transformer, possibly noising all around. And as we're actively regulating anyway, there's no benefit from reducing ripple from very small to extremely small. Would make sense (perhaps) in full-unregulated supply.

rule-of-thumb :1000uF per Ampere - make do with a single 2200 or 4700uF

I don't like the way the +68V is generated from the +250 - a transistor fail might be catastrophic (depending on what it's used for)

Has anyone an idea for a power transformer that could deliver 120mA @ 225 - 250V and 4A @ 12V ?

You will want to talk to Tomasz Lachowski from TOROIDY.PL - Ask for a quotation, tell him you need it for sensitive audio, and he'll dimension it at selected cores, low core saturation, magnetic AND electrostatic shielding. Say hello from me :) - I'm a long time fan of their work

If you're heating at 12VDC, you'll probably want some 14-15VAC from transformer

/Jakob E.
 
Last edited:
Dear Jakob,

I don't like the way the +68V is generated from the +250 - a transistor fail might be catastrophic (depending on what it's used for)

Do you mean to better have a drop with a resistor before? Or better having a dedicated supply voltage (winding) to generate the +68V?

Thanks again for the valuable feedback. I appreciate your advice.
Falk
 
Last edited:
I looked into the schematic again, and it's safe enough as-is - there's no reason to change. The 68V is just for varimu screen grid, so it don't mess with the real world

( reason for dislike is that I've seen +48V for phantom derived like this some times, and it horrifies me..)

/Jakob E.
 
Last edited:
Dear Gentlemen,

its been a while. During the past 10 month I barely could afford to work on the project aside from maintenance and calibration here and there. After one year working with the compressor on an almost daily base I dont want to miss it anymore. This motivates me to go on with the project and cure its downsides. I finally found time to:

- redesign the PSU from linear to regulated
- redesign the filament supply from AC to DC
- order parts
- order new power transformers - Thanks Jacob, I spoke with Tomasz. What a great guy!
- design the PSU PCB
- redesign the audio circuit PCB

I still having trouble to get the JJ EF806S matched and will try the TAD EF86 instead soon. If anyone could provide me with a matched sextet of any of both types I would appreciate an offer. Which kind of tube tester would you recommend to buy?

I attach the schematic of the new PCB. If anyone finds something that would not work - please let me know.
Thanks in advance,
Falk
 

Attachments

  • Bildschirmfoto 2021-10-09 um 11.06.45.png
    Bildschirmfoto 2021-10-09 um 11.06.45.png
    315.8 KB · Views: 53
  • Bildschirmfoto 2021-10-09 um 11.08.32.png
    Bildschirmfoto 2021-10-09 um 11.08.32.png
    385.1 KB · Views: 46
Regards tube tester - I can highly recommend uTracer. It's DIY, not expensive and controlled from PC with a lot of parameters that can be changed and tested.
Regards tubes - you can try Tung-Sol EF806SG, they are new and I think they should have pretty tight tolerance. Around 30euro each.
 
..happy to learn that you interface well with Tomasz at Toroidy - such people can be worth their weight in gold

- the JJ "quality control" and "matching" is a joke. So is, unfortunately, their general quality IMO. I would stay away, like entirely away, from their products.

Your problem with the EF86/EF806 is that they were adopted by the hifi crew, which exploded prices and completely crashed quality. If it was me, I would look closely into the slightly-different-heater-version, the PF86. This one is still widely available in real 1975-quality because it's luckily unusable for 6.3V hifiheating. Hint: the "P" indicates it to be a 300mA series-heater type, the data sheet mentions 4.5V heater voltage, so for each tube we need to drop 1.8V at 300mA (isn't that ~ 6 Ohm /0.6W in series with each heater?). I think I have a large-ish box of these somewhere, let me know if I should send you a handful to try out

/Jakob E.
 
TAD offers matching. They have a selfmade tubetester which measures an entire batch of tubes.
Other than that I´m sure that we can talk our member MrDude into matching a bunch of tubes. He has one of these testers: Röhrenprüfgerät Röhrenmessgerät Roe - modern computerized tube tester

Falk, if you decide to go the PF86 route shoot me a pm, I might have some of these left over.
 
I don't understand why you would need any other tubes than the remote-cutoff tubes to be matched in U23. The output tube is a cathode follower and the two tubes before it are inside a negative feedback loop and have a operating level fine adjustment trimmer. What am I missing here?
 
Back
Top