Modifying the M552 / M553

[frank23]

My setup has evolved to such a level that the M553 I use as a crossover has by far become the weakest link in the setup. Therefore I decided to examine the circuit, do measurements and see what mods could be made to improve the soundquality. One could say that there are a great number of other active crossovers available, but none of those had what I was looking for out of the box like the M553 has. As the M552 circuit is exactly the same save for the 2-way vs 3-way, the below also goes for the M552.

In the circuit are about 32 capacitors in the signal route. As in high end systems the sound of even one capacitor is deemed critical to the overall system sound, you can image what 32 can do to the sound. But if all these had a function, upgrading them would be quite a lot of work and cost quite some euro's, although the time involved would be much more of a burden for me.

So I analysed the circuit and made measurements with multimeter and scope to see what happened after each opamp. It turned out the DC was quite low at all points in the circuit. Capacitors are normally used to prevent potentiometers from crackling by blocking the DC. But here the DC is only in the area of 40mV max. This could still be a problem if the amplification was large, but in all the steps in the M553 process, there is about unity gain with all pots set at 0dB. So the ratio between DC and signal was always about the same.

Therefore I thought it would have to be possible to bridge 26 out of the 32 capacitors. Incredible, 26 out of 32 could go. And I just did it by bridging all the 26 pairs of legs by small soldered jumper leads of about 5mm. And the M553 now sounds awesome. Depth, air, body, mellow, lucid, name it. It can finally keep up with the rest of my system.

So, why did JBL include all those capacitors in the first place? I think for predictability in how the M553 would function. With the capacitors it is more foolproof, less of a worry I guess, as the potmeters can't crackle, there is no plop when muting etc. But in reality, in my living room, the 26 could go.

6 remain as they are the output capacitors and are the final block for DC leaving the M553 (both hot and cold of the balanced outputs have one each).

So, if you use the M553 in a high end setup (or any home setup I'd say), bridge the 26 caps and free your sound!

ps., and see to it that the subsonic filter is bypassed by setting the jumpers the right way...

[just4kinks]

Can you be specific about which of the capacitors you removed?

I have glanced at the schematics and I don't see any unnecessary capacitors. I hope you're not talking about the filters.

I would suggest buying some 1% polypropylene film caps, like the MKP1837. And if you don't adjust the frequency often, you could replace the pots with fixed 1% metal film resistors.

[hjames]

Wow and wow! Very cool discussion ... thanks for sharing this info!


6 remain as they are the output capacitors and are the final block for DC leaving the M553
(both hot and cold of the balanced outputs have one each)

JBL 553 Specs & Schematic (http://www.jblproservice.com/pdf/M-Series%20Electronics/M553.pdf)


My setup has evolved to such a level that the M553 I use as a crossover has by far become the weakest link in the setup. Therefore I decided to examine the circuit, do measurements and see what mods could be made to improve the soundquality. One could say that there are a great number of other active crossovers available, but none of those had what I was looking for out of the box like the M553 has. As the M552 circuit is exactly the same save for the 2-way vs 3-way, the below also goes for the M552.

In the circuit are about 32 capacitors in the signal route. As in high end systems the sound of even one capacitor is deemed critical to the overall system sound, you can image what 32 can do to the sound. But if all these had a function, upgrading them would be quite a lot of work and cost quite some euro's, although the time involved would be much more of a burden for me.

So I analyzed the circuit and made measurements with multimeter and scope to see what happened after each opamp. It turned out the DC was quite low at all points in the circuit. Capacitors are normally used to prevent potentiometers from crackling by blocking the DC. But here the DC is only in the area of 40mV max. This could still be a problem if the amplification was large, but in all the steps in the M553 process, there is about unity gain with all pots set at 0dB. So the ratio between DC and signal was always about the same.

Therefore I thought it would have to be possible to bridge 26 out of the 32 capacitors. Incredible, 26 out of 32 could go. And I just did it by bridging all the 26 pairs of legs by small soldered jumper leads of about 5mm. And the M553 now sounds awesome. Depth, air, body, mellow, lucid, name it. It can finally keep up with the rest of my system.

So, why did JBL include all those capacitors in the first place? I think for predictability in how the M553 would function. With the capacitors it is more foolproof, less of a worry I guess, as the potimeters can't crackle, there is no plop when muting etc. But in reality, in my living room, the 26 could go.

6 remain as they are the output capacitors and are the final block for DC leaving the M553 (both hot and cold of the balanced outputs have one each).

So, if you use the M553 in a high end setup (or any home setup I'd say), bridge the 26 caps and free your sound!

ps., and see to it that the subsonic filter is bypassed by setting the jumpers the right way...

[jblbgw_man]

I concur with just4kinks … most of the “coupling” caps are forming part of the filter sections ….. if you have removed them it might be worth doing a frequency sweep to see what you have ended up with, sounds like you could be surprised or disappointed:blink: … Traditionally in Opamps, coupling caps are used to reduce the effects of DC Bias offsets on the front end of the Opamps from the preceding Opamp stage, the voltages are just a few mV or lower and will vary greatly with temperature. Now imagine a series of cascaded Opamps one after the other all with a DC Offset, the first Opamp multiply it’s DC Offset by its gain and then follow through on all the Opamps multiplying by their gain … by the end you could quite easily end up with offsets of some many volts in either polarity … all subject to temperature drift, now superimpose an AC waveform on that DC Bias of say 10 volts …. Then this is going to reduce the overall AC level you can use (by the DC Bias level) as you approach the supply rail voltages and all unstable and temperature dependent and it goes without saying this will also severely reduce your headroom as well …. Capacitors are NEVER used for “to prevent potentiometers from crackling by blocking the DC” :eek:… if you want to stick with the M552 and improve the sound then replace C9, 91, 97, 98, 40 and 41 with Elna Cerafine caps or some other audio grade cap …. These will be the main culprits and if you want to continue then replace all the signal caps with polypropylene film caps or so …. And contemplate doing a Opamp upgrade ….. personally I would be removing all those wire links you have soldered in and your speakers might appreciate it as well! .. good luck !!:scoot:

[frank23]

1) of course I did not remove the caps that form the filter section
2) the coupling caps I bridged are not part of the filter section
3) I checked that all gain stages, from output of the previous to the output of the next, are all unity gain, so no large buildup of DC will take place, otherwise that would indeed have been a problem
4) I didn't know DC was temperature dependent, I'll check it once it has been warmed up

I'll post the numbers I bridged tonight, but they are the caps around the pots and mute buttons, and some between sections.

try it!

[subwoof]

a few things:

Remember that DC offsets can go either polarity / amount and if many are in series, may just cancel out the overall amount.

Also by changing the type of op amp to a non-compensated version, and adding a trim pot, you can bring the offset to essentially zero. this is standard form for small signal amplification in labs.

And the outputs can also be changed to true diff output chips that do not use series caps and have great CMMR - rane uses these and many console manufacturers.



And I happen to have a pair of the 553's in the "stash" ....maybe worth looking into for the big house speaker system. wasn't too keen on going digital..:)

sub

[Ruediger]

See the 4th book on this page: http://www.douglas-self.com/ampins/ampins.htm

Ruediger

[frank23]

Ok, the caps I bridged are: C19, C96, C28 (I think, can't quite decipher the PDF, 15uF NP before the first NE5534), C21, C39, C97, C98, C42, C101, C102, C45, C104, C105

This is for channel 1, the equivalent caps for channel 2 are also bridged, making 26 in total.

I have left on the modded M553 overnight and it still works fine. I did notice that the nuts that fasten the potentiometers to the front have to be fastened as otherwise when touching some of them a slight hum was noticable. I left them of yesterday and left the nuts and knobs off, but when firing up the system I noticed a slight hum when touching the shafts so I put the nuts on again and fastened them and that made it silent again.

The change in sound quality is still amazing. Whole new worlds open as the rest of my system was so much better, the M553 can now keep up again. And yes, I could have exchanged all the caps, but the best cap is still no cap in my experience. Now I still need a few black gates in the power supply and in the output.

[frank23]

Just wanted to let you know the M553 stil plays without fault, so now I'm sure all is well. And the sound quality benefit is very large indeed. It made me sell me sell my beloved high-end cd player as suddenly it was clear another combination I had was much more open, and the M553 finally had the quality to show it. Before, it levelled all outputs to the same mediocre quality, now differences are clearly audible.

[4313B]

See the 4th book on this page: http://www.douglas-self.com/ampins/ampins.htm

RuedigerPublication Date: June 24, 2011

I missed that one, thanks for the heads up! My copy shows up this Friday.

[Guido]

Hi Frank

Thank you for bringing this up.
Certainly DC-coupling of hifi equipment brings dramatic improvements in sound. The reason you don't find this in commercial equipment is because it is dangerous if you don't know exactly what you do. I've seen you do but one need to have some electronic skills to run a fully DC coupled chain. That's why commercial amps/preamps go the safe route and block DC at inputs and outputs of gain stages.

I run a Pass X2.5 type DIY preamp based on the UGS 4 design without any caps. The improvement when removing the caps was almost shocking! BUT the DC-Offset must be carefully controlled.

Next step is to remove output caps from the DAC which can be problematic as it is not so DC stable as the X2.5

To remove the input caps from the Mono Amps (XA100.5) is only for FAB (Fearless Amplifier Builders) :) With a Gain of 26dB even a small DC offset can cause serious destruction.

[frank23]

To remove the input caps from the Mono Amps (XA100.5) is only for FAB (Fearless Amplifier Builders) :) With a Gain of 26dB even a small DC offset can cause serious destruction.

And as I run my 2420 without any blocking cap between amp and drivers, this would not work for me :-)

I need amps that are reliable and have low DC offset. I use the very nice Marantz MA24 monoblocks designed by Ken Ishiwata. I have 6 of them, 3 for each channel l/m/h. They have a built in DC offset monitoring thingie.