Need help with a crackling ML Spire

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi bolserst,

Thanks a million for all your advice :) I'll get those speakers disassembled right away again (got enough practice by now, I dare say ;-) and see if my local electronics shop has these on stock as they did with the resistors. Will hopefully be able to report back tomorrow evening... fingers crossed!

Cheers,
André

P.S.: just noticed that the caps you linked to are actually really good with 10% tolerance. Most others only have a 20% tolerance. I may end buying those if my electronics dealer doesn't have anything comparable
 
Last edited:
Just a quick update: I didn't make it to town to the electronics shop today, but I did resolder all points on the board and that did not remedy the crackling either, so I believe I can safely exclude a bad soldering point.

I also measured the diodes and caps in the HV section includ the R2 resistor (onboard though) with a component tester and all measured ok, which is puzzling. I'm still considering to replace all with new parts just for the sake of it.

Something keeps me thinking about that square transistor in the Royer Oscillator section or one of the mosfets bolted to the heatsinks. What are the chances that it's one of these?
 
...Something keeps me thinking about that square transistor in the Royer Oscillator section or one of the mosfets bolted to the heatsinks. What are the chances that it's one of these?
The mosfets are the active switches used in the limiter circuit.
Test 3) from post#18 bypasses these mosfets.
If bypassing doesn't help, they(and their control circuits) are not the problem.

Your description of the slow but repetitive noise still sounds very much like something on the HV side of the Royer oscillator to me.
I would replace the diodes first, test. If still crackling, replace the caps and test.
 
Haven't done the bypass test of the limiter circuit yet and still need to. This won't damage the esl stators, will it?
No, not unless you have a really powerful amplifier and turn the volume way up.

The limiter threshold is set at an audio input level ~ 40Vrms. (200W/8ohm amplifier).
Above that level, the limiter starts working and keeps the voltage level getting to the ESL stators below the threshold for corona/ozone generation even if you are driving it with 600W amplifier.
 
Could not it be some whisker like structure or other sharp spike developed on PCB surface due to flux residue? It's fare to say that dirt cheap HV ceramic caps are prone to failure without a reason and do have high piezo effect i.e. can emit the sounds if discharge happens inside.
Is it possible to run board without the panel connected and see if the cracking persists?
P.S. Once a discharge on the surface occurs even once it will create path of least resistance which may not be real carbonization = easily seen. May be, just may be if nothing else helps proper PCB cleaning with with aerosol flux remover and clean layer of acrylic HV Conformal Coating will do?
 
That was my initial suspicion too, but the ML boards are very well soldered with no whiskers anywhere and no carbonization visible anywhere either. They are additionally coated with a clear protective lacquer in the HV section that should protect from surface discharges. I've also been coating my new solderings with a coat of "plastic" spray for electrical circuits as well, so I still think the fault is rather to be found in one of the components than in surface discharge (apart from the fact that I inspected the board at night during power-up and couldn't see any blue sparks).

I was at our local electronics shop today but the did not have either the UF4007 diodes (only 1N4007, which I'm concerned will be too slow) or the HV ceramic caps in the right voltage, so I've ordered them in the bay and it'll be a couple days before I can report any update... please bear with me :)
 
Last edited:
I can hear a type of what I call ... Mylar expanding.. you could call it cracking.. when the panels first gets hit with Bias Voltage .....on my ML, JansZen, Acoustat Audiostatic ,Quad...it last only 3-5 sec then it stops...but ML is the only one that turns bias off an on..not great for getting the best sound out of a ESL panel...well to my ears...I have a pr of ML SL3s setup for the bias to stay on all the time ...best sound I ever got out of the ML...thanks to Bolserst help
I have always heard this .....it sounds like the one board is letting the bias leak off the panel more on one panel when it turns off... I would guess
good luck on getting it to stop
 
Last edited:
Short update and sorry for the long break... I was off on a bisiness trip for a week and am still waiting for the ceramic caps. The guy from ebay tried to fool me by selling me 10% tolerance caps and sending me standard ones, so am awaiting the replacement any day. Hope to report the outcome on the weekend. Please stay tuned :)
 
So, it looks like the Ebay vendor is trying to jip me - no response and I'm tired of waiting and am going to buy the ceramic caps at Farnell now. I'm going to get MuRata caps, but noticed they have different series, i.e.

A) the DEH series (125 deg.C) with reduced heat dissipation due to small dielectric loss of the ceramic material and suited for use on high frequency pulse circuits such as a horizontal resonance circuit for CTV and snubber circuits for switching power supplies.... or

B) the DEB series (85 deg.C) for use on decoupling circuits for power supplies.

See here for more infos: http://www.farnell.com/datasheets/5769.pdf?_ga=1.57521708.78757811.1478987195

Which do you recommend that I take?
 
…my Spires are driven by a Unison Research Performance tube amp with 40W per channel and for the tests I always open up volume slowly and just enough to activate the speakers.
Sounds like you have a good handle on careful testing procedures. :up:

…The guy from ebay tried to fool me by selling me 10% tolerance caps and sending me standard ones
Sorry to hear about the difficulties with ebay seller :( With 100% positive feedback on their record you would think they should be a trusted seller. In this application, the exact value is not that important +/-20 to +/-50% is not a problem. As long as they are made by reputable manufacturer and rated for 2kV or more they will work just fine.

… I'm going to get MuRata caps, but noticed they have different series, Which do you recommend that I take?
Either will work fine in this application. I have used both series in my HV supplies with no failures to date.
The pin width on the DEB looks like it will match the PC board better, so I would probably go with those.

Attached is a schematic I traced out for the oscillator and HV multiplier portion of the board.
You may find it useful during your troubleshooting.
 

Attachments

  • ML_Spire_HV_circuit.PNG
    ML_Spire_HV_circuit.PNG
    38.5 KB · Views: 113
Dear bolserst,

Thanks for your reply :) And don't worry about the Ebay seller, I'm chalking this off as bad luck. In my hurry to get the caps ordered and to be on the safe side I went with the higher spec'd DEH... didn't want to take any more risks ;-)

Even with the diodes I went for ones from Vishay. Those have already been soldered in. Just need the caps now, which should be here probably by Wednesday as Farnell has been quite reliable over here. Can't wait to get my Spires up and running again... I miss the music *deepsigh*

While we're waiting and out of my incessant curiosity :D... does it pay off to replace the big caps on the crossover-board containing the coils with better ones? If yes, which would you recommend that fit into the cramped space in the speaker? I did already replace the Jamicon electrolytics on all boards (except the ICEPower boards, which I left untouched) with Elna Silmic II and one or two Nichicon KZ which made a slight but noticeable improvement. I would have loved to replace the big foil caps with Jantzen Superior Z-Caps, but, alas, they only have up to 22uF and are huge.

I did read the other thread on the CLX, where mods to the Spire were discussed, but moving components into another cabinet is out of the question for me momentarily. I'd like to see, what can be done in the stock housing for now.
 
Let me put my previous question back for now... first things first and unfortunately bad news: neither the replacement of the caps nor the diodes helped. The crackling noise is still there, so I will next try the bridging of the series pass portion of the limiter... *deepsigh*, but I won't give up till I've got this bugger fixed.
 
Dear Friends, I think I may have found the error thanks to a tip from another member of the forum. He indicated that he had discovered similar crackling in ML panels if the output voltage was too high and mentioned that he had occasionally found the resistors R211, R214 and R217 to fail under the permanent high voltage running over these, causing this failure. I unsoldered a leg of each and low and behold, R217 measures as defect part in my component tester, so this may be the culprit. I won't be able to tell for sure until I have replaced them (I'll replace all three on both panels as prophylactic measure) ... need to order these.

@bolserst: do you happen to know the rating of these resistors? I'm not sure with the third and fifth ring, if these are red or brown as the colour is so similar to the first two rings, which I read as red. I assume it is brown, which would mean these are 2.21m Ohm 1%... could that be correct?

Thanks in advance :)
 
A faulty component finally located. This is good news!

Yes, they are 2.21Mohm 1% 1/4W 500V resistors.
I had thought a couple times about having you check those, but the symptoms just did not seem to point to them, so I didn't. :eek:

These resistors form part of the voltage divider used for feeding the sensed output voltage back to the voltage regulator portion of the circuit. During normal operation, they will see roughly 1200 VDC across the three, so 400 volts across each. This is running pretty close to their voltage rating. Since there is room on the board, you might consider replacing them with 1/2W 700V resistors.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.