New development on old project: Roger Mayer RM58

[MidnightArrakis]

Hi
As you noted I have no basic knowledge... Learning slowly from my errors
It s just a miracle I can simulate and put together components for my diy projects
Thanks a lot for making me notice what s wrong!
Of course I d like to read something more about pcb design...
Can you upload your files in tech Doc... I believe should be useful also for other people who want to learn and investigate....
Thanks for your attention
Best

I'm assuming that your PCB layout was done using "KiCAD". If so.....could you send me -- BOTH -- your KiCAD schematics and PCB layouts? That way, I could directly edit your layout and then send it back to you and from there, you could learn by looking at the differences between your and my layout of the exact same thing. Sound like a plan to you? Let me know, OK???

Meanwhile.....here is the link to an entire folder of "PCB Design Reference Materials" that you can download and learn from:

https://mega.nz/folder/6uRS0bqZ#6OoLf3qKN2F474FcSM2JdQ

ENJOY!!!

/

 

[richiyobs]

Great! Thank you so much....
Best

 

[Spencerleehorton]

I’ve got a load of these built from a 3d printer which I’m using for my version of this.

 

[richiyobs]

It s Rm58 ...
I used an old opa version for the moment
Due to Lorlin size now i have to redraw a new Gr led Meter to fit in the front panel

 

[thomasdf]

I am looking for 4 2N3820 matched in Europe, or if someone has an alternative FET reference that's close enough and easily found I'm interested !
I still have leftover PCBs for the comp and amp too.
Cheers & thanks
Thomas

 

[richiyobs]

I am looking for 4 2N3820 matched in Europe, or if someone has an alternative FET reference that's close enough and easily found I'm interested !
I still have leftover PCBs for the comp and amp too.
Cheers & thanks
Thomas

Banzai music ...
but not so cheap!!
https://www.banzaimusic.com/search.php?mode=search&page=1&keep_https=yes
2n5461 costs less


they can match parts with extra money i believe

 

[thomasdf]

yes they can ! thanks for the ref !

 

[thomasdf]

Hello all !

Quick update on my build : so far I have completed one channel. Very happy with how it turned out, and the sound it seems to deliver.
The trimmer are indeed a bit mysterious but with some time and an oscilloscope and the documents I think I've ended up with something I am pleased with. Can't speak for the meter now, but for the audio it is working.

The tricky part is RM57 and RM58 do not have exactly the same pots value (esp timings!) and hookup arrangement between comp board and the makeup amp board, so one need to pay attention to that, and the documentation I have is a little unclear. Earlier in this thread, someone posted docs for the RM58 and this is correct, and must be followed if building the RM58 with soapfoot boards like I did.

My only concern now is that the input pot (100K log) is a bit "everything or nothing" for the moment, volume wise.
I have an unbalanced input and 1:1 Sowter balanced output. My output amp distorts the signal quite a bit, and I still need to decide if I like it that way or not

Will report back later ! Cheers

 

[thomasdf]

I have bought a pair of 100uA meter ... and they are completely unusable and I can't set up the trims to have them display something useful or even legit. They peg like crazy... So far the best result I had was a 1mA meter...

 

[Ivan K.]

Regarding your input pot, you can always add a resistor across the whole pot to change the overall output, or change the taper by adding a resistor from the wiper to either the high or low side of the pot. Many other configurations are possible using more resistors, but simple is usually good enough!

 

[thomasdf]

I am curious to know if I’m the only one with this issue, since I used the same value as in the schematics and manual ..
I will have to find a solution indeed, it is not very usable as is.

 

[richiyobs]

I have bought a pair of 100uA meter ... and they are completely unusable and I can't set up the trims to have them display something useful or even legit. They peg like crazy... So far the best result I had was a 1mA meter...

try to lower the resistor at the matched Jfet drain... also try to alter the resistors at the base of bjt, part of the meter driver
if you can simulate you will find the right amperage for your needs. New development on old project: Roger Mayer RM58 thake a look here at that part ... hope it helps
i m using both 1mA TN73 type and 100uA-PM2
anyway as you said using RM58 values for release, 10Meg and 100k for attack, your tracking needle will move better.... i fought a lot with that meter driver, i m not an egineer, only fun with electronic.... but the oveshoot is due to Jfet characteristic, especially with fast attack
once I even thought there was an extra circuit for that wild meter... but nothing found

best

 

[thomasdf]

Thanks for your feedback ! The schematic is a bit cryptic for me, where is the sidechain input ? I'd like to completely disconnect the sidechain in order to do some testing... It feels to me like it's a feedback type compressor, and sidechain is fed the output, near pin L on the schematics, before the output pot to makeup amp.

 

[richiyobs]

It feels to me like it's a feedback type compressor, and sidechain is fed the output, near pin L on the schematics, before the output pot to makeup amp.

You re right.... I can t Watch at the original schemo now on mobile... So i don t rember what pin L Is... If i don t go wrong the sidechain input has a 10uf cap followed by 100k trimmer to set the threshold point
Best

 

[thomasdf]

I am finishing up my RM58, and for some reason ... it sounded really good out of the rack with a 15VDC SMPS maxxed out to approx 16/17VDC.
No everything is racked up, grounded, etc, and I have a similar Meanwell 24VDC SMPS delivering 20VDC and I get some high freq noise from the main board. Output amp is clean. I am a bit puzzled ! With batteries the unit is dead silent, so it def has to do with that meanwell SMPS, and they are usually dead silent too, which is why I use them..

 

[Ivan K.]

Do you have a second MW supply to try? Otherwise, try adding an Electrolytic cap AND a 0.01uF Ceramic right at the output screw terminals. If there is room, you could solder the 0.01uF to the bottom of the PCB at the output traces. If you have a SMD, .01 to 1 uF cap and can solder it to the output, even better. I'd try anything from 220uF to 1000uF for the electrolytic.

 

[thomasdf]

i tried that with no success ... I have another MW SMPS that I tried, same result only a tad lower in volume, but maybe it's because it's less powerful than the original one. I am a bit puzzled because I've never had problems with Meanwell SMPS in many other builds. I have tried using a 100uH choke, nothing changes. I have also noticed than during startup my output pot is scratchy like if it had DC, then it's fine. It is brand new...
With a 7918 regulator after the SMPS it is much quieter. Maybe I'll go this route with a 7920

 

[Ivan K.]

It could be a ground issue. I'd check that the MW output common has close to 0 Ohms to the case and also 0 Ohms to your RM58 case and common lead. Also, if you want to try a really low noise linear regulator, the LT3042 should do it. Available as a pre-built module on eBay. You could also try ferrite beads on the output lead. Loop the output wire through the bead several times.

 

[thomasdf]

So I finally went SMPS to 7918 regulator to power rail. Output of the regulator as an extra 1000uF 50V capacitor. It is dead quiet.
I need to get into calibration (tomorrow probably, it's already late here) but it already sounds really good. I am impressed... Input is unbalanced, output has Sowter 600:600 transformers.
I had extra noise too because the mu metal cases of the transformers were not properly grounded.
I'll probably re do all the grounding tomorrow. Thanks for your help Ivan ! I'll report back when it's calibrated

 

[thomasdf]

well I won’t lie, calibrating this thing is not easy and is a bit of trial an error.

I followed @richiyobs advice, as well as @soapfoot advice too. R27 is a 50k trim, and has a lot of effect on the comp.Tweak a symmetrical waveform with this. Much easier to do with a scope and square wave generator than a sine. Then, with the input/threshold knob at a fixed position, set the signal generator to get 0dB at input. Play with VR6 and VR5 to get symmetrical waveform and same level at input and output of the main board. We don’t care about the makeup amp during calibration. That’s it for the audio path / signal detection.
For the meter, it is much more mystical but I was able to get decent results by setting up VR1 and VR2 alternatively, going back and forth on one and then the other so that the meter rests at 0dB GR. channels must be unlinked and release set to slowest.
It helps to be able to (true)bypass the channel so the meters sits in an « idle » position and then compare this idle position with 0 GR.
VR1 was used to fine tune the position when everything is set up.
My meters are 1mA. They are not super accurate for low GR (strange ballistics) but they are spot on at -10 and -15 dB and that’s close enough for me.

I then realized the compression kicked in a little too early (too low thresh) and too hard for my taste so I relaxed VR5 a bit

Everything was done with a 600 ohm load at the output, first with a scope and then with Protools and some audio. Then readjusted and double checked with the scope again before closing up the rack enclosure. I am using 100k pots for the attack instead of 22k, so I’ll probably change that in the future.

Thanks a ton @soapfoot and @richiyobs , can’t wait to use it in a real context. Very happy with the sound.