Technical Assistance/Advice Thread

[bhjazz]

Pandemic Project #54: re-case the Ventus.  So, some quick history: I won the raffle Nate had back in the day for the TwistedPear Ventus.  It has worked fine and I have enjoyed it for years but I wanted it in a different chassis.  Over the years I bought parts here and there.  The case came from VT4C.  I had top and bottom panels made with venting and finally got it all reinstalled recently, only to set up a vicious game of chase the ground loop noise for myself.  The setup is, otherwise, the same as the original setup with all the same parts, just transplanted into a different case.  The only difference in parts was changing the power button setup (which had a nice relay) to a bog standard clicky switch.  The boards next to the amp boards just have a cap (with a bypass cap).  Nate added these to tame some of the output voltage drift, if I remember correctly.  

So the noise through the headphones sounds like 120Hz ground noise, but more like a sawtooth wave than a sine wave.  The amount of noise varies as the volume is turned up - it starts off loud at 0%, then goes away until about 25% volume, where it gets loud again, then  fades until you hit full volume, where it is again loud. 

Touching the chassis changes the noise slightly, and touching the output ground also changes the noise.  If I turn the volume up to the "silent mode" at about 50-60% I can touch the ground and bring the noise right back.  

I posted a similar question on the TwistedPear forums, but received pretty much nothing but questions. When I answered those, the discussion went dark.  Thanks, TP audio.

I have attempted various fixes that did not work:

• Disconnect the signal ground connection from the chassis ground (see pics below)
• Disconnect the ground bridge between the L and R channels on the Alps volume pot
• Change the ground connection from the left channel board to the right channel board
• Connect the grounds together on the amp boards
• Connect the ground from the Alps volume pot to the chassis
• Remove the Alps from the chassis
• Cursing and shouting at the noise

I don't have a 'scope to check the incoming voltage from the power supply, but do wonder if something is amiss there.  

Any guesses for chasing the ground loop dragon?  I really hate to see this amp gathering dust.  Damn.  Thanks for any suggestions.  

BH

[n_maher]

Crazy, I thought about that amp recently and couldn't remember where it went. One thing, I have NO memory of adding those boards/caps and they certainly looking like output coupling capacitors. The Ventus had/has(?) a history of DC on the output IIRC.  I don't remember all that well, actually, but I've never see PMBs like that, that I can remember anyway.

One thing to check would be the influence of the transformer, since it's both unshielded and in the same box with the amplifier boards.  I'd try very carefully rotating it while you're listening to the noise and see if you can find a position when the noise gets better. I remember chasing something like this on the first M3 that built and in the end the only thing that fixed it was aggressively shielding the toroid. 

[mwl168]

I built a Ventus EZ (Ventus was out of stock) years ago for my father-in-law using very similar set up and it was hum and noise free. IIRC, there are pots on the amp boards to zero-out the output offset and I did not need to add coupling caps to the output. 

Hard to tell what's going on from the pictures. Looks like you shorted the inverted input to the ground on both channels but why do you have two ground wires coming out of one board and none for the other?

May be check the wiring and ground scheme? 

[dsavitsk]

It's worth checking the transformer, but that seems like the less likely issue here. Transformer induced noise will generally either increase with volume (if it is being picked up on the input wires) or be constant with volume (if it being picked up past the pot). And you have pretty good distance here.

The grounding, however, does not look very good. It is difficult to see exactly what is going on, but what you need is a single ground point that everything connects to directly. I usually put two terminals somewhere, such as in the middle of the case, spaced about an inch apart. Run a thick-ish piece of wire between them, and call that your star. Connect everything - the PCBs, the jacks, the RCAs, the pot, the PS, directly to it.

For earth, connect the earth pin on the IEC to the case as close to the IEC as possible. If everything (jack, pot, RCA) is isolated from the case, then you can use a ground loop breaker to connect that point to your ground star, but if you are grounded correctly, it is probably not necessary, or useful. Instead, I'd just connect that point to your star ground, too.

Edited May 31, 2020 by dsavitsk

[Pars]

2 hours ago, bhjazz said:

The case came from VT4C. 

Is VT4C still around? Their website has been down for quite some time.

Is the pot itself grounded to the chassis (thru the nose)? Other than that, what Doug said makes good sense.

[bhjazz]

Wow, you guys are great.  I'm so glad I could come back here and get such straight answers.  Many, many thanks.  

 

3 hours ago, n_maher said:

Crazy, I thought about that amp recently and couldn't remember where it went. One thing, I have NO memory of adding those boards/caps and they certainly looking like output coupling capacitors. The Ventus had/has(?) a history of DC on the output IIRC.  I don't remember all that well, actually, but I've never see PMBs like that, that I can remember anyway.

One thing to check would be the influence of the transformer, since it's both unshielded and in the same box with the amplifier boards.  I'd try very carefully rotating it while you're listening to the noise and see if you can find a position when the noise gets better. I remember chasing something like this on the first M3 that built and in the end the only thing that fixed it was aggressively shielding the toroid. 

Nate: you did not have the boards there.  You simply had the caps hanging off of the amp boards.  I wanted to keep the output caps from being so fragile, so added the boards (from Jameco) and the bypass caps since they were staring at me.  I have wired the two boards together underneath.  

I'll try the "transformer twist" first.  Thanks!

 

3 hours ago, mwl168 said:

I built a Ventus EZ (Ventus was out of stock) years ago for my father-in-law using very similar set up and it was hum and noise free. IIRC, there are pots on the amp boards to zero-out the output offset and I did not need to add coupling caps to the output. 

Hard to tell what's going on from the pictures. Looks like you shorted the inverted input to the ground on both channels but why do you have two ground wires coming out of one board and none for the other?

May be check the wiring and ground scheme? 

Ah, I did not check the offset.  Will search for the pots as well, just in case.  Thanks!

The ground scheme is what was in the original build: one input ground, one output ground.  Connecting these two to the corresponding spots on the other channel did not work for me.  

 

3 hours ago, dsavitsk said:

It's worth checking the transformer, but that seems like the less likely issue here. Transformer induced noise will generally either increase with volume (if it is being picked up on the input wires) or be constant with volume (if it being picked up past the pot). And you have pretty good distance here.

The grounding, however, does not look very good. It is difficult to see exactly what is going on, but what you need is a single ground point that everything connects to directly. I usually put two terminals somewhere, such as in the middle of the case, spaced about an inch apart. Run a thick-ish piece of wire between them, and call that your star. Connect everything - the PCBs, the jacks, the RCAs, the pot, the PS, directly to it.

For earth, connect the earth pin on the IEC to the case as close to the IEC as possible. If everything (jack, pot, RCA) is isolated from the case, then you can use a ground loop breaker to connect that point to your ground star, but if you are grounded correctly, it is probably not necessary, or useful. Instead, I'd just connect that point to your star ground, too.

Very interesting.  I see two different schemes here: Connect (ground) everything or if everything is isolated then use some type of ground loop break.  I think mine is a case of the ground noise volume being constant with volume, for the most part.  Not sure why it comes and goes with volume change...  I do feel like this chassis has totally different metal conductivity, so that may have turned into a problem when I tried to duplicate the original wiring.  

How would I connect the PCBs and the PS to this new ground?  Using the ground off of the board itself, or the ground it is "processing"?  That is, with the power supply for instance, if I check continuity between the screw that holds the board down and the chassis I get nothing.  (My wording may be convoluted here.  Blame the Syrah.)

I do find it odd that if I connect my DMM to the incoming ground connection (at the IEC) and any random point on the chassis bottom that I get no continuity.  However, if I do the same test and use the rear panel there is clearly some connection.  I'll post what I find with some better testing.  Thanks for leading me in that direction.  

 

2 hours ago, Pars said:

Is VT4C still around? Their website has been down for quite some time.

Is the pot itself grounded to the chassis (thru the nose)? Other than that, what Doug said makes good sense.

No, VT4C is long gone.  I think they closed in 2010, so now you know how long ago I started buying parts!  

Now that I have tested a little bit today, the pot is grounded to the front panel, but that may not be at equal ground potential to the rest of the build.   (?)

 

Again, many thanks to all of you.  I'll try these, test a bit more, and report back.  

 

Edited May 31, 2020 by bhjazz

[Craig Sawyers]

It isn't clear if the various bits of the new case are electrically connected. Many case manufacturers powder coat or anodize all the parts, and leads to odd problems with ground continuity.

[bhjazz]

You're definitely on to a piece of the puzzle, Craig.  I actually can't get a connection to check resistance on the front/rear panel unless I have the probe inside a screw terminal.  

Yesterday I tried the trafo twist and was able to knock the noise down just a tick.  Cool.

I also found that the weird volume pot noise (noise at about 50% volume) was more due to unshielded headphone cable location (the ESW9 that I had on) than anything else.  So now I'm just hearing it at 0 - 25% on the volume pot.  This is pretty close to what I hear with a pair of Grados. 

I started measuring everything to give myself a better picture.  Here are some values.  BTW, my Fluke 112 shows 0.2 ohm even with the leads shorted together.  

First, these are taken with the amp OFF.  

across caps inline from earth to rca grounds: 10.2
earth to rca gnd 10.2
rear panel chassis to rca terminal ground: 10.2
rear panel chassis to output ground (on amp board) 10.2

earth to ps board screws (I'm guessing this is the ground plane?): 0.2
earth to amp board screws: 0.2
earth to bolt holding down trafo:  0.2
power switch case to power supply screw term:  0.2

Channel A (top board, left channel) 
Channel B (bottom board, right channel)
input gnd to output gnd (channel A) 0.2
input gnd to output gnd (channel A) 0.2

earth to chassis (front faceplate)  0.2
alps pot chassis to output gnd  10.2 
front panel to rear panel 0.2

earth to output ground 0.2
earth to input ground 0.2

But here is something I found really odd: with the amp ON:

Alps chassis to output ground on the board that *does not* have an output ground connection - 10.2 and immediate silence through headphones
Alps chassis to output ground, on the board that *does* have  an output ground connection - 10.2 and huge buzz through headphones

This is heard most clearly with a pair of ESW9a 'phones. 
 

I'm going to start by bypassing the pair of caps that connect earth to the input RCA jack ground.  That would eliminate most of the 10.2 ohm readings above.  

Next, I'll connect the ground screw from the pot to the current star ground, and connect input grounds together and output grounds together.  We'll see where it stands at that point, but the tests with the amp on make me wonder what else is going on.  

Thanks!

 

 

[bhjazz]

I had tried some of these connections before, but adding them together was the key.  The amp has been dead quiet for an evening or two.  More soon. 

[n_maher]

What happened to the old chassis? It was one of the few that I modified on my benchtop milling machine when I had it?  I remember precision drilling all of those holes...

[bhjazz]

It's around here somewhere, I believe.  The dent in the lower portion of the face plate was kindof getting to me...  Probably happened during shipping here back in the day.   

So I ended up doing just a few things and it all locked in.  I connected input grounds on the boards together, as well as the output grounds.  This was weird to do since the left and right grounds were already connected at the Alps.  Finally, I connected the ground from the output jack to the ground plane screw on the closest amp board.  

DEAD QUIET.  Finally!

Hey, many thanks to you Nate, Doug, Craig, and Pars for all the ideas.  I'm glad to have this one back in the arsenal!

 

Will post some pics now that it's done.  Still looking for a 3mm green LED for the new front panel, though.  (I think it was a 3mm)  I have no idea why that item is so hard to find these days.  

Edited June 7, 2020 by bhjazz

[dsavitsk]

On 6/6/2020 at 9:44 PM, bhjazz said:

.  Still looking for a 3mm green LED for the new front panel, though.  (I think it was a 3mm)  I have no idea why that item is so hard to find these days.  

https://www.mouser.com/ProductDetail/Vishay-Semiconductors/TLLG4400-AS12?qs=ZTK%2Bjoi3Sg6xGMde4j6A1A%3D%3D

 

[bhjazz]

Thanks!

[tomislavkufrin]

For a simple balanced volume control, is attenuator + 4 diamond buffers OK, or i misunderstood something?

I saw that people were making preamps with input selection and ubal/bal boards, but i suppose if i only need balanced, ubal/bal board is optional?

[jose]

If the input signal is balanced or you don't want an extra gain on the input, you don't need ubal/bal boards.

What diamond buffers are you going to use?

[tomislavkufrin]

I saw three different gerber files. 1 with trough-hole, 2 with smd components. The one i was looking at was "diffinput6a" - with smd components. 

I saw your project also, with trough-hole components, so that is also an option.

[iemphile]

I'm looking for books to read to better understand amplifier design. I'm mostly interested in solid state dynamic and electrostatic headphone amplifiers. I have taken some EE courses in college, but those were mostly theoretical or focused around digital devices and had little actual analog circuit design. 

[jose]

Douglas Self "Audio Power Amplifier Desing"  

ISBN-10 : 0240526139

 

It is not complete but it will help you understand some things.

Also look for Nelson Pass articles (they are available on the web) and READ all threads where Dr. Gilmore, spritzer, etc ...

 

[chinsettawong]

I really like to try to make some wire stators for my DIY electrostatic headphones.  I've bought some wire mesh, but I have no idea how to stretch it so that it's absolutely flat and tight.  Any ideas?

 

[dncnexus]

Hello all, 

I was making this post to try and get some advice/help on troubleshooting/repairing my Eddie Current Electra estat amp. 

Background: This amp is a balanced design, with a separate left and right channel circuit. I got this from Australia and this amp has a PSU that you can jumper to the proper voltage of your country. I have swapped the voltage to the 110V slots with the jumpers.

Problem: The problem is upon startup, I get volume out of both channels. After approximately 45 minutes - 1 hour, I will get the right channel to cut out and go silent. This can happen by either the right channel distorting then cutting out, or it sometimes just immediately cuts out.

Stuff I have checked/Additional Information: 

- This amp is inherently balanced but has XLR and SE inputs. I only have a SE DAC so I haven't tested the balanced inputs. 

- I have checked the internals of the amp to see if I could find any components that didn't look right. I found 2 solder joints off of the right channel PSU intake that were white and ashey. I have since resoldered these joints but the problem persists. 

- When the channel cuts out, if I unplug and replug the stax plug but at an angle where the right channel pins touch first, I am able to get sound from the right channel, but upon full plug in, the right channel stops again.

- There are 2 stax pro bias jacks. When the right channel goes out, both of the jacks exhibit this, not just one. Looking inside, they are wired separately and don't utilize a jumper between the two jacks. 

- This is consistent across 2 different headphones, and when running the headphones on another Stax SRD7 energizer, they work fine with no problems with the right channel.

- I have swapped the tubes from the left and right channel to see if it could be a tube problem. It was persistent on the right channel

Things I was going to try to troubleshoot:

Since this is a balanced amp with dedicated right and left channel circuits, I was going to go through and compare each part of the path to see if the left and right align in their readings and to see where things start to differ to isolate the problem.

What I am looking for:

I am looking for additional eyes/minds on what this could be, whether there are ideas on what to check or if anybody has experienced a problem like this before. I am not the most educated/knowledgeable on amps/repairs so was just looking for guidance, or ideas on what this could be so I can try and repair. If I really can't figure it out on my own, I will try and reach out or find someone to assist/help fix this, but wanted to learn and try this on my own. 

I can photograph and upload any photos people need of either the amp or the psu or both. 

Thank you for all the help to anybody who reads or gives help/ideas.

Cheers and Happy Halloween 

 

 

[spritzer]

Can you post some detailed pics so we can better understand what might be going on?  If something stops working after a while like that than it is usually a component, solder joint etc. loosing contact due to thermal expansion but it could also be something else. 

[dncnexus]

11 hours ago, spritzer said:

Can you post some detailed pics so we can better understand what might be going on?  If something stops working after a while like that than it is usually a component, solder joint etc. loosing contact due to thermal expansion but it could also be something else. 

Yes, when I get home from work I will take detailed photos inside.

[dncnexus]

Here is an imgur thread with the internal pictures of the amp and PSU. I can take closer pictures of sections that need closer detail. 

Thank you

https://imgur.com/a/pDetzXR

[kevin gilmore]

in the past, craig had trouble with all those white plastic connectors with poor crimp on the pins. look carefully for browning or burning on these connectors, removing them one at a time and reseating them.

[dncnexus]

2 hours ago, kevin gilmore said:

in the past, craig had trouble with all those white plastic connectors with poor crimp on the pins. look carefully for browning or burning on these connectors, removing them one at a time and reseating them.

I ended up reseating all of the connectors and noticed the right channel rca cable connector wasn't seating very nicely, so I cut and soldered the wires to the appropriate pins. I have tried out the amp again and it still does the same with the right channel cutting out about 45 mins into the listening, and will intermittently come back after disconnecting/reconnecting/changing the connection. 

I will probably try to trace those jack wires back to see where they are and check out that connector/area to see if anything looks odd there.