Op-amp mod for DOD R-825 Compressor

canakas

New member
Significant mods for the DOD R-825 Compressor/Limiter

In reference to the post quoted under I wish make a previously mentioned topic resurface, for the benefit of anyone who wishes to sprite up their mono compressor;

Daniel Gomez said:
The DOD R-825 absolutely features the dbx 2150A VCA chip. In fact it has two. One on the regular compressor circuit and one in the de-esser section. I just opened mine up last night to check on it as I am a "modder". Everthing is in sockets so another great replacement would be to upgrade the 5532 opamps to AD826, a much faster (slew rate) and wider bandwith device. I will be selling mine, is any body interested? I used it on Bass alot as it was pretty transparent. Not a bade compressor but not very good for vocals although the Mod would help with that. I need to cull some stuff I don't use any more from the racks. Good luck. :)
You can see the schematics here

Now, I wonder which one is the 5532 opamp, as I cannot read that number anywhere on the layout.

Also, I would like to call on you for your suggestions for a replacement opamp.
As you can see the quoted poster suggests thithe AD826 (is this the one?)

Why would the AD826 make the R-825 better on vocals?

And what do the terms opamp bandwidth and speed mean - What are the effects of variations in these?

An lastly, is this an easy mod? Is the opamp socket the same?


All your replies are greatly appreciated :D

best regards
Canakas

EDIT: 30.04 Changed the thread title to fit the content.
 
Last edited:
The RC 4558 is the equivalent of the 5532 - the 5532 has slightly better specs for slew rate and rise time (giving a greater bandwidth/frequency response.)
http://www.fairchildsemi.com/ds/KA/KA4558.pdf
http://www.ortodoxism.ro/datasheets2/5/0q263do8iq991qftdprcsxqcjsky.pdf

The ADA 826 betters these parameters by an order of magnitude, in theory improving the frequency response of the unit. They all have the same pinout so if they are socketed you should be able to interchange the DIP forms of the devices.
http://www.analog.com/UploadedFiles/Data_Sheets/AD826.pdf

Sometimes increasing the bandwidth will introduce unwanted instabilities into a circuit such as oscillation, so there's no absolute guarantee that the switch would work (or that it would even improve the perceived output signal from the device.)

Good luck - it's worth a try as the parts are cheap and the switch seems to be simple.
 
If swapping out op-amps is all your comfortable with then keep in mind you don't need to change all of them. Only the ones carrying your audio signal will matter for the upgrade.
As Crazydoc mentioned you need to watch for spurious oscillation with higher bandwith devices. The oscillation frequency can be in the Megahertz for a device like the AD826 so you will not hear it. Only an oscilloscope will show this. This type of oscillation will eat up most of the bandwidth of the opamp.

I've found that upgrading resistors and power supply usually makes a more dramatic difference in percieved sound quality than changing opamps.

If you are comfortable with soldering I suggest replacing all the cheap carbon resistors in the audio path with high tolerance metal film types which have a much lower self-noise.
Then you can consider turning your attention to the power supply circuit. The device uses the 78xx and 79xx regulators which aren't all that good on noise rejection. A big improvement can be made here by swapping the regulators for LM317 and LM337 which have improved noise immunity and better regulation tolerance. You will have to improvise a little room for the adjustment circuit these devices require though. A benefit of adjustable regulators is the ability to trim the voltages so your power rails are symmetrical. Symmetrical power rails will eliminate a great deal of DC offset at the outputs of all the opamps and improve performance overall.
The stock regulators have a 4% regulation tolerance which means the power supply rails may be unbalanced by 8% or more. The resulting DC offset this will create in the opamps not only hurts audio performance but also in the precision rectifier circuits that generate the control voltage for the VCA. Eliminating offsets here improve the dynamic range of the VCA and the accuracy of the controls associated with it.

To really improve this unit for vocal processing I would suggest not using the De-esser feature and use the old sidechain EQ method. The de-esser circuit here is a not a good design at all. It uses a 2 pole fixed frequency state variable filter to provide both high-pass and lo-pass filtering to split the signal like a crossover. The high frequency component signal and the low frequency component signal are then compressed separately by the two VCAs and then recombined. The problem here is that whenever you combine the HP and LP outputs of SVF you wind up with a notch filter response. Oberheim used this in their old analog synthesizers where it made for a great effect. In a compressor though, you just wind up with a hole in your sound. Not very good for vocals especially if your vocalist is a soprano.
 
The board has plenty of space

top of the mornin and thanks for the replies,

The pcb is spatious and the ICs are all in sockets, so changing them is a piece of cake. I count six of them in the signal path, not counting the one in the de-esser(not too pleased with it either really)

As for oscillation measurements, there are equipment available at uni.
Now if there is oscillation, would it help to better the power supply and thus the rail symmetry?

Also, would there be any point in changing the LF353's that are in the signal path as well? I do not understand enough of electronics to see what their function is... What would I change them for?

As for the power supply, theres is plenty of room to include some more stuff... how big is the regulator circuitry by the the way... an inch by an inch maybe?
By the way, is there a standard design for the LM317/LM337 voltage? I feel I have seen them before, with the boxes with the double screws, right?
Where can I get a schematic layout for this?

Starting on the power supply is smart, because in the same job I can make the compressor able to feed on 220V instead of 120V so I wont have to use a step-down transformer.

Changing the resistors is somewhat of a bigger job, and my soldering skills are ehem in process of becoming acceptable :o
It would make a big difference, I agree with you on that.

From what I can see the ones on the board are all carbons.
What kind of soldering tin is recommended for audio applications?
Is it the ones with silver in them?

A final entry, where does one buy these components? I have used a place called futurlec once, but that was for a non-sensitive application(fan control).
What dealers do you recommend?


Thanks again.
 
canakas said:
As for oscillation measurements, there are equipment available at uni.
Now if there is oscillation, would it help to better the power supply and thus the rail symmetry?

I don't think so. Killing oscillations usually involves either changing the ground path of the device or band limiting it with a small value capacitor from it's negative input to its output.

Also, would there be any point in changing the LF353's that are in the signal path as well? I do not understand enough of electronics to see what their function is... What would I change them for?

For an older opamp the LF353 is still a capable unit for most audio apps. In fact I use it often in my own circuit designs. If I was to upgrade or replace it I would use another BiFET type device such as the AD712JN.

From the schematic you have one 353 acting as a differential amp on the balanced input. Another right after it as an adjustable gain amp. The next is after the VCA and I suspect it's only purpose is impedance buffering. If you are running balanced signals into your compressor then the 1st 353 on the input would be the one i would NOT replace. In this role the 353 has excellent CMRR (common mode rejection ratio).


As for the power supply, theres is plenty of room to include some more stuff... how big is the regulator circuitry by the the way... an inch by an inch maybe?
By the way, is there a standard design for the LM317/LM337 voltage? I feel I have seen them before, with the boxes with the double screws, right?
Where can I get a schematic layout for this?

Look at this datasheet:

http://www.national.com/ds/LM/LM117.pdf

you want the circuit bottom of page 15. Use a 25 turn trimmer for the adjustment so you can tweak the voltage tight.
Here's the DS for the negative regulator:
http://www.national.com/ds/LM/LM137.pdf

top left schematic on page 6 is exactly what I would use to replace the power supply!

These regulators come in the same size package as the ones in the compressor now. Watch the pinouts though, you cannot just drop the adjustable regulators into the holes for the old fixed regulator devices.

Starting on the power supply is smart, because in the same job I can make the compressor able to feed on 220V instead of 120V so I wont have to use a step-down transformer.

I don't see anything in the schematic that indicates a high voltage primary on the power transformer. Play it safe and stick with your step-down unit.

Changing the resistors is somewhat of a bigger job, and my soldering skills are ehem in process of becoming acceptable :o
It would make a big difference, I agree with you on that.

From what I can see the ones on the board are all carbons.
What kind of soldering tin is recommended for audio applications?
Is it the ones with silver in them?

A final entry, where does one buy these components? I have used a place called futurlec once, but that was for a non-sensitive application(fan control).
What dealers do you recommend?


Thanks again.

In the UK try Farnell. For USA use Mouser or Allied. Absolutely do not order from Futurelec again unless you like the possibility of being ripped off. Had to send my credit card company after them about a year ago.

The resistors are easy to change and there really aren't that many in the signal path if you don't count the de-essing circuit. Solder is up to your preference. I don't recommend the new lead-free kind for this work.

btw: if you decide to change the resistors be careful because the 1% metal film types have a different color coding scheme than the old carbon resistors. If you aren't sure then use a DMM to measure the resistance of each as you go.
 
Last edited:
I sure have some work ahead

Cheers mate, for your sound answers!

I will post back when I have completed the work. Or when I am stuck more likely :D

Have a nice 1. of May!
 
The DOD-825 reopened...

Hi people,

I feel really bad about not having completed the mod I was asking you questions about years ago, but my chemistry masters degree completely consumed my time... Now that it is over (!), I can finally return to the audio section, and to getting my mono compressor a notch up and running.

I have decided to switch some of the RC4558Ns with AD826ANs and some of the LF353s with OPA2132PAs.

I am about to order and I feel I need to confirm what parts I need. I will leave the de-esser alone (I don't use it).

Could someone who is more proficient with schematics confirm the involved parts in the main audio path (cf. schematics in post #1):

U1 - LF353: no change (as Somnium7 explains in post #5)

U11 - LF353: -> OPA2132PA (before the dbx2150A)
U9 - LF353: -> OPA2132PA (after the dbx2150A)

U10 - RC4558N: -> AD826AN (output gain stage?)

U7 - LF353: What does it do? Is this even an audio application?

U12 - LF353: What does it do? -> OPA2132PA
U13 - LF353: What does it do? -> OPA2132PA

Others not to change:

U2 - RC4558N: no change (de-esser)
U3 - RC4558N: no change (de-esser)
U4 - RC4558N: no change (de-esser)
U5 - RC4558N: no change (de-esser)
U6 - RC4558N: no change (de-esser)

U8 - LM324N: no change (LED control right?)

I would very much appreciate your help on this! :D

best regards
 
Half of U7 is a buffer for the control voltage; the other half is a LED driver. U10 is the output driver. Be careful swapping that as it is configurated such that inverting attenuation is possible, and some opamps don't like less than unity gain. U12 and U13 are the sidechain; they take the signal, rectify it, and filter it to a control voltage. They don't need to be particularly sexy parts to do that job, since it's likely they are much faster than your attack or release settings.
 
Hi, thanks for the reply and for sharing!

This means I should probably leave all U7, U12 and U13 alone.

Do you however think it is worth hoping that the AD826 is stable with less than unity gain? (what does this really mean?)
I see that the data sheet says it is unity gain stable, but I am not sure how this reflects on its stability with less than unity gain...

Also, does it seem reasonable to change the U11/U9 - LF353s for OPA2132s? Are these the most influential OAs in the main sound chain in is this design?

Thanks again.
 
I don't know that I would bother. Most of the sound will come from the VCA chip anyway, and it will be the limiting factor for distortion and noise.
 
I suppose I should trust your intuition, but I'm really curious to hear what happens or not =)

What are the chances of damaging the unit?
The suggested opamps have the same pinouts and voltage operating areas...

What are the symptoms of oscillation, should it occur?
 
I suppose I should trust your intuition, but I'm really curious to hear what happens or not =)

What are the chances of damaging the unit?
The suggested opamps have the same pinouts and voltage operating areas...

What are the symptoms of oscillation, should it occur?

Low, those chips have fairly similar specs.
 
I'll post back when I've tried them out.

And now I do so... although it has been a long time!

When I opened the R-825 I saw that one of the LF353s present in the schematic at U9/U11 (cannot recall which) was indeed a RC4558N. So I played around a little which changing the three opamps U9/U11 and U10 around with the RC4558N, LF353, AD826AN and OPA2132.

Since I am not an experienced compressor user, I tried to listen for clarity, dynamic range and noise. I ended up switching all three for AD826AN, a configuration that is slightly bassier in the sound than combinations of 2x RC4558N/1x LF353, and much the same as 2x AD826ANs and 1x LF353 on U9 (I chose the AD826AN here for consistency with the U11 OA before the dbx chip).

The Burr-Browns on U10/U11 were good too, but with a slightly fuzzier on high volumes and not so 'wide' sounding as AD826AN.

If the unit is oscillating or if this really is a better setup than stock, I cannot tell. When I play with the compression, it is not completely transparent, but it doesnt make the sound less interesting either. For some uses the OPA2132s might have been more interesting choice.

I'll keep this setup for a while, and post back if I find somethings not right.

If anyone has questions or suggestions, I would be happy to hear them! :D

-canakas
 
Back
Top