Acoustat bias power supply snubbers

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This thread is about one of those very rewarding things that a person wishes that they could have done many years ago. The improvement in sound quality that I found just by adding three basic parts to my speakers has been both unexpected and amazing!

Over the last few years I have been modifying my audio equipment with great success by adding optimised snubber networks to the power transformers to eliminate ringing, using a Quasimodo device designed by Mark Johnson:

https://www.diyaudio.com/forums/pow...sformer-snubber-using-quasimodo-test-jig.html

Installing these snubbers has resulted in substantial improvements in the sound and has been a very inexpensive and relatively easy process. After I had performed this procedure on the rest of my audio system components, I decided to try to design and install a similar snubber network for the transformers in my Acoustat high voltage bias power supplies. The details of that are posted below, but first some important disclaimers.

*****Warning! High voltages present inside Acoustat interfaces can injure or be fatal, and may still be present long after power has been removed from the device. Only work on or modify Acoustat interfaces if you have experience with high voltage electronics and are familiar with proper safety procedures. Minimum spacing and parts positioning awareness is required to install these modifications. If you are unfamiliar with this type of electronics, find someone who has the qualifications to do the work for you.*****

*****These values are only applicable to the Acoustat MK-121 and MK-131 five diode multiplier interfaces with the 750 volt transformer. Older interfaces with the three diode multiplier would require different components, and this modification is not compatible at all with the wall-wart type power supplies.*****

OK, now on to the details. Mark Johnson's snubber design method involves determining the optimum values of a three element network consisting of two capacitors and one resistor. I recommend that everyone interested in performing this upgrade to read the information presented in the above thread, particularly the first post and it's attachments.

Using my Quasimodo device, I first determined that for the Acoustat power transformer the optimum values for a snubber would be as follows; Cx 10nF, Cs 150nF, and Rs 4.7Kohms. Since the power transformer outputs a nominal 750 volts AC, high voltage capacitors would be required. I measured the transformer output to confirm the actual voltage, and found it to be quite close to 750VAC at 120 VAC input.

Now, the voltage markings on capacitors actually indicate their rated DC voltage and in many cases the AC voltage rating is much lower than that, so the datasheets must be consulted to determine if the capacitor can safely withstand the applied voltage. Unfortunately there are very few choices for capacitors rated high enough for this application, and I was unable to find a suitable 150nF capacitor for Cs. I did, however, find a 100nF capacitor with a sufficient voltage rating:

B32656S2104J564 EPCOS / TDK | Mouser Canada

So my choice at this point would be to use two of these parts in parallel, or to see if perhaps 100nF alone would be sufficient. There is not a lot of room inside my interfaces to add extra components, and I found that installing two of these capacitors would be very challenging. I then did some more testing with the Quasimodo and determined to my satisfaction that 100nF would perform satisfactorily in this case, so I opted to go ahead with just one 100nF capacitor for Cs.

For Cx I found this part:

PHE450XD5100JD15R06L2 KEMET | Mouser Canada

Rated at 1000VAC.

For Rs I first calculated the power that the resistor would be required to handle, and came up with a value of 3.64 watts. Since I already had on hand some 4.7K 10W resistors, I used them for this project. Mark recommends a carbon composition or metal film resistor for Rs, but I have not searched for a suitable part myself and so am not sure what is available. Using a resistor greater than 10 watts is recommended to reduce heat stress, but this would depend on what is available and space considerations. I definitely wouldn't advise using anything less than 10 watts in this case.

I went ahead and carefully installed these parts into the interfaces, making sure that there was sufficient clearance between each part and connection to not cause arcing, and let the speakers charge back up again before conducting listening tests. I was not expecting huge changes in the sound with this modification, but both my wife and I were very pleasantly surprised at the differences that we found. There was a big reduction in grain and harshness to the sound, everything sounded smoother but much more detailed. I found myself hearing fine details in the music that I had never noticed before, and found it to be much more involving. Instruments were better defined and more firmly located in the soundstage, and overall the sound was much more natural and musical. The extent of the improvements was greater than many other more expensive and complex modifications that I have made to my system over the years, and was an intensively satisfying upgrade.

I don't know if having previously eliminated transformer ringing in the rest of my system made this final upgrade more effective, or whether the changes would be so dramatic otherwise. I would be very interested if someone with the abilities and resources to do this upgrade could confirm the improvements on their own speakers. I have attached an updated schematic showing where to install the snubber components in the power supply. Please feel free to post your comments and questions!

Take care,
Doug
 

Attachments

  • acoustatsnub.gif
    8 KB · Views: 182
Thanks for posting anything About Acoustats…An getting better Sound....
The High V bias is a place I have found were sound inprovements can be had on any ESL ever made!
But back to the Acoustat MK 121 interfaces Bias....In the pics I have posted of my interface,an as minny have seen before.... there are 2ea 1/4 watt fuses in each leg of the primary input to the bias treansfourmer….

Andy or others can say what the fuses are there for...like life saveing….
I have never seen this size 3-5 amp, 750 v bias transfourmer used on any other ESLs an it dose add to the Sound of the Acoustat ...or any other ESL panels type... I have used these treanfourmers on...allway better Base...

But Any thing done to these fuse....putting 1/2 watt in place of the 1/4 watts
Just diff type, or just running one fuse an a wire on the other leg like I seen done......well-can change the sound!
These small fuses seem to help the high frequency response!
So anything like your Snubber, or new bias parts in the bias it self....like a diff type 500meg feeder res....diff caps an diodes....changes the sound ...most of the time for the better.

All just mans finding
 

Attachments

  • images[1].jpg
    images[1].jpg
    7.8 KB · Views: 449
  • images[7].jpg
    images[7].jpg
    10.5 KB · Views: 442
Add 2*100k Ohm resistors on the secondary wires. specially before you ground the transformer. Maybe it will destroy your quasimodo cirquit. As a rule i always try to increase the impedance in the "beginning" of the power supply and decrease the impedance close to the user.
 
Add 2*100k Ohm resistors on the secondary wires. specially before you ground the transformer. Maybe it will destroy your quasimodo cirquit. As a rule i always try to increase the impedance in the "beginning" of the power supply and decrease the impedance close to the user.

Transformer snubbers must always be installed as close to the transformer secondary wires as possible, with no parts in between. Adding resistors after the snubbers will not affect the performance of the snubber itself, but I don't understand how doing so would improve the performance of the bias supply or the speaker. I don't believe that the impedance of the power supply is at all affected by the snubber. Can you elaborate on this modification and it's effect on the circuit?

Take care,
Doug
 
In todays Audio,most of us dont wont anything biger than a cell phone...
But these Big Acoustats,Its like trying to get out of a 30 year love afair...well for anyway!

I just pic up a all Stock 40 years old Conrad Johnson PV 5,it sounds just like a Tone Magic box on the 1+1s or 2+2s M 3
CLSs lot fun

Puting the two interfaces bias on the same Variac, let you fin tune the sound..

Thanks for any input on getting good-better sound out of any ESL setup
 
Last edited:
I agree that the snubber should be optimized to create a "resistive" power source, and that it should be placed as close as possible to the secondary winding. This makes the tranformer "ideal". Adding series resistors is like adding dampers in a mechanical system. It dampens oscillations that is generated by the rectifier diodes.
I always adds resistors to the rectifier. And it reduces not only galvanic carried noise but also radiated noise ( EMI ).

Hooking up a scope to a power supply with a small current shunt, measuring the current spikes is a good lesson.
The lower impedance and more fancy big and many RIFA PEH 200 lytics the more EMI Disturbances.
So your own power supply use to be the worst noise generator.
That is my point.
 
Thank you both for your comments. I have no experience with any other electrostatic speaker bias power supplies other than the Acoustat type, so I cannot comment on them.

I have also not experimented with increasing the "impedance" of the Acoustat power supply to reduce noise, but I would be concerned that those additional 100K resistors could limit the current supplied to the voltage multiplier, negatively affecting it's operation. Perhaps the lower current would limit the size of the rectifier spikes and their conduction back to the transformer, which might reduce transformer ringing to some degree, but they certainly would not eliminate it completely. However, they would also be changing the resonant frequency of the transformer inductance/rectifier capacitance network, possibly making the problem worse.

A properly designed snubber will completely eliminate the transformer ringing without having to make any other modifications to the circuit that might negatively affect it's operation. Adding resistors between the transformer snubber and the rectifiers may compromise the snubber's effectiveness and I don't recommend doing that at all. I don't believe that the rectifier spikes themselves are a significant problem once the transformer ringing is dealt with, just like with lower voltage power supplies in other audio equipment. If you are concerned about it, adding a filter capacitor to the output of the voltage multiplier as others have previously suggested should completely eliminate any rectifier spikes from reaching the speaker diaphragm itself.

Making other modifications to the bias power supply may or may not result in improvements, but from my experience I believe that eliminating transformer ringing has the largest sonic benefits. I strongly recommend installing the optimised snubber network first, and then experimenting with other changes to the power supply to see if they can further improve it's performance.

Take care,
Doug
 
Last edited:
Hi, just out of curiosity,
your measured voltage of 750v ac - would you not need to work out the peak to peak voltage and then size your capacitor voltage accordingly?
As I said, just curious,
regards john

Hi John,

The capacitors that I used have both DC and AC RMS specifications and are rated above 750 VAC with a much higher DC volt rating. Voltage ratings are defined on page 6 of this document:

https://www.mouser.ca/pdfdocs/HowtoOrderKEMETFilmCaps.pdf

Take care,
Doug
 
cap voltage

Hi Doug,
As I say I was just curious because our 240v mains has a peak 2 peak voltage of 340v which caused the early demise of some 240v movs. After researching I found this also applies to capacitors. Not that long ago I bought some "reputable brand" capacitors of the correct class as mains filters, and in every case they all caught fire destroying the amplifiers, and almost setting the house on fire. I emailed the company, who wanted to know which parts were used, and I heard nothing since. I presume they were fearful of being sued. But these caps were rated for across the uk mains and advertised as self healing! Clearly this was not the case. Obviously when I hear about people using caps like this it sends alarm bells ringing.
Thanks and regards - john
 
Hi John,

I'm not quite sure what message you are trying to send here, that manufacturers can't be trusted to supply parts that meet their own specifications, or that working with mains powered equipment is too dangerous for DIY, or is it just general paranoia based on your one bad experience?

It sounds like the manufacturer of your capacitors sold you some defective or incorrectly specced products. They failed while being used for their intended purpose, which is certainly not your fault, and I would recommend that you not ever consider using parts from that manufacturer again. Perhaps if you were willing to share the name of the manufacturer it could serve as a warning to others that their parts can be defective and should be avoided. I would also be questioning the mains fusing in your amplifiers if it failed to protect the device from an internal short circuit.

The Cx capacitor that I used is specified for AC and pulse use, and is rated at 3000VDC/1000VAC, which should be fine on a 750VAC transformer.

The Cs capacitor is rated at 2000VDC/800VAC and is designed and specified for use as a snubber, which is exactly what the application is here. As it is in series with Rs, it will not even see the full voltage output of the transformer.

If one is concerned with parts failures, a higher rated part could be used, or alternatively, four capacitors in a series-parallel arrangement would double the effective voltage rating (space permitting.)

The line fuses on a 120VAC Acoustat interface are rated at 1/4 amp, which will allow a maximum of around 30 watts of power. Therefore the 750 volt transformer rating must be something less than that. Installing capacitors across the secondary is much different than a capacitor placed directly across the mains input, since the available current is limited by the transformer.

As I said, I'm not quite sure what your point is here, as I am using parts for their intended purpose, within their specifications. That this would "set off alarm bells" makes no sense to me, and implies that manufacturers and their products cannot be trusted. If it is because you are not confident in working with voltages above 50V or so, then like I stated in my initial post, you should not be doing this type of work at all.

Take care,
Doug
 
As I say I was just curious because our 240v mains has a peak 2 peak voltage of 340v which caused the early demise of some 240v movs.

MOVs are rated in both DC voltage and AC RMS voltage, NOT peak voltage. If you installed a 240VDC rated device on your 240VAC mains then it is no surprise that it failed, and if it was a 240VAC rated MOV then you did not take into account line overvoltage conditions and under-rated the MOV. Here is a typical MOV application guide that recommends a 140VAC RMS rated MOV to be appropriate to install on a 120VAC line; therefore you should have used 280VAC RMS rated parts for your application:

https://m.littelfuse.com/~/media/el...ng_a_littelfuse_varistor_application_note.pdf

After researching I found this also applies to capacitors.

This is absolutely not true. As I have stated above and shown in the linked information sheet, capacitors intended for use on AC lines are rated for AC RMS volts. Yes, the peak voltage of a 240VAC line is higher than the normally used RMS value, but so would the voltage rating of the capacitor be higher in peak volts. You are comparing apples and oranges here.

Take care,
Doug
 
I perused the 2013 thread and can't say as I understand what problem a snubber is supposed to fix. Don't the high frequency oscillations get filtered out downstream? Or have no audible consequences?

What measurable improvements can be demonstrated with a snubber installed?

If somebody could post a few words of Snubber 101, I'd appreciate that.

B.
 
Hi Ben,

My understanding is that the high frequency oscillations are radiated into the air and conducted back into the AC mains, causing various issues with whatever equipment that they encounter, including the device that they are installed in. They certainly have easily audible consequences as attested to by my own and many, many other people's experiences, on this site as well as other audio and general electronics sites. Using transformer snubbers has been a standard industry practice for many years.

I personally have installed optimised snubbers on a number of pieces of audio gear, including an Audio Research SP-9 MKII preamplifier, PS Audio power amplifiers, various home built amplifiers and preamplifiers, as well as the Acoustat speaker interfaces that are the subject of this thread. In every case there was a substantial improvement in the sound quality of the equipment as described in my initial post.

Unfortunately I don't have access to any precision distortion measuring equipment, so I have no "paper" data to support these observations. I suggest that you repost your questions to the Quasimodo thread where literally hundreds of well-respected fellow DIYers have found the advantages of optimised snubbers to be both compelling and worthwhile.

Take care,
Doug
 
Hi Ben,

My understanding is that the high frequency oscillations are radiated into the air and conducted back into the AC mains, causing various issues with whatever equipment that they encounter, including the device that they are installed in.

Got the concept. Many thanks.

But isn't there an easy A-B test that could be done blind? YouTube has a bunch of snubber videos, but I couldn't tell which might have a test demo and with audio gear.

B.
 
Got the concept. Many thanks.

But isn't there an easy A-B test that could be done blind? YouTube has a bunch of snubber videos, but I couldn't tell which might have a test demo and with audio gear.

B.

I'm not sure how an easy A-B test would be structured, as the snubbers cannot be quickly and easily removed or reinstalled. If a person had two identical pieces of equipment, one with snubbers and one without, it might be possible but you would still need to quickly switch between them and power down the one that was not currently in use. This situation is similar to the benefits claimed by using soft-recovery diodes in power supplies, do you know if that modification has ever been measured or tested in this way?

I do think that you should post your queries in the Quasimodo thread, where there are people who are better qualified to answer your questions.

Take care,
Doug
 
I'm not sure how an easy A-B test would be structured, as the snubbers cannot be quickly and easily removed or reinstalled. ...I do think that you should post your queries in the Quasimodo thread, where there are people who are better qualified to answer your questions.

No disrespect as we Canadians like to say, but you are the one advocating use of snubbers here so maybe you shouldn't just brush off my request for supporting evidence just like that.

And if my nose is still functioning well, remarks that you don't know how to test something everybody can hear, don't smell right.

B.
 
No disrespect as we Canadians like to say, but you are the one advocating use of snubbers here so maybe you shouldn't just brush off my request for supporting evidence just like that.

And if my nose is still functioning well, remarks that you don't know how to test something everybody can hear, don't smell right.

B.

My intent with this thread was just to share my positive experiences with using what I have learned from others in this forum. You have asked questions that I don't have the answers to, so I directed you to ask them to the people who might. I don't see that as brushing off but rather as being helpful and I'm sorry that you don't see it that way. And I really don't understand what my personal testing abilities have to do with anything.

Take care,
Doug
 
lets see he (ben) does not know how to measure this but he expects you to explain it to him (and to provide proof) to his liking or you are some sort of troll. funny how that works. it just goes on and on this "I don't believe you but you need to (try to) prove it to me anyway" no win no way. It's such total crap.

If someone advocates what sounds like miracle cures for unsubstantiated problems* which are here to fore not known to the unwashed masses (not to mention ignored by virtually all high-end manufacturers too), it is not unreasonable to expect them to justify themselves (even just a little bit) if asked for background information and not to rudely blow-off questioners.

This ain't the Gwenyth Paltrow webstore.

Maybe my system needs snubbers and maybe it doesn't. I sure don't know yet. i think, moray james, if you read the dialog again, you won't be quite so judgmental.

B.
*which they claim are destroying sound quality all over the place although passing unnoticed by the rest of us or to quote OP: "very rewarding things that a person wishes that they could have done many years ago"
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.