SMPS repair tip

[dave]

On 07/28/2014 11:01 PM, Phil Allison wrote:

"Phil Allison"

uses only four TO220 Mosfets.

There are no class D amp chips and the mode of operation of the output stage
is mysterious.



.... Phil

What is a Tripath amp?

 

[Cydrome Leader]

Phil Allison <phil_a@tpg.com.au> wrote:

"Cydrome Leader"
Phil Allison

** The lower the resistor's value, the more significant inductance
becomes.

Egs:

I use a pair of 100W rated, 8 ohm, tubular, wire-wound resistors as a
dummy
load - wired in series for 16 ohms and in parallel for 4 ohms. The
ceramic
tube is about 1 inch dia and 4.5 inches long with 34 turns of flat strip
conductor would along it, so the inductance is about 6uH.

At 20kHz, the error caused by inductance is under 0.5% while at 100kHz,
the
error is still only 10%.

For the 1 ohm wire-wound sense resistor, the inductance was about 200nH.
However the ringing frequency was around 3MHz making the inductive
reactance
almost 4 ohms - four times more than the value of the resistor !!!

well, it's known that giant ceramic wire wound power resistors are not the
choice for RF use.


** Sure - but this is a current sense resistor in SMPS circuit where most
of the current is at 100kHz but with a strong harmonic at about 30 times
that frequency. Problem being, the over-current trip was being activated
prematurely because the peak value of that harmonic was being exaggerated by
a factor of 3.

I agree with this- for the power supply problem- who'd have expected this
to happen?

While 1W metal film resistors have low enough inductance to work OK in the
circuit, typical wirewound types do not.

BTW:

I could have added a RC network ( ie a Zobel ) across the 1 ohm wirewound to
bring it line at 3MHz and well beyond. A 47nF cap and any 1 ohm, 0.5W film
type would do fine.

Same goes for the 100W 8 ohm power resistor, a Zobel using 68nF and 8.2 ohms
corrects it to beyond 5MHz if that was ever needed.

I wonder how much that would confuse the next repair tech?

Bizarro repairs are always fascinating to come across.

 

[dave]

On 07/30/2014 03:31 PM, Cydrome Leader wrote:

dave <ricketzz@earthlink.net> wrote:
On 07/27/2014 11:57 AM, Cydrome Leader wrote:
dave <ricketzz@earthlink.net> wrote:



I hope you used special gold plated, cyogenic magical copper cables for
all that.

No. 14g stranded usually. 12g for the woofs.

 

[dave]

On 07/30/2014 10:59 PM, Michael A. Terrell wrote:

Cydrome Leader wrote:

well, it's known that giant ceramic wire wound power resistors are not the
choice for RF use.


Carborundum resistors are used for RF.

My dummy load is full of tab resistors clamped to aluminum heat sinks.

 

[Michael A. Terrell]

dave wrote:

On 07/30/2014 10:59 PM, Michael A. Terrell wrote:

Cydrome Leader wrote:

well, it's known that giant ceramic wire wound power resistors are not the
choice for RF use.


Carborundum resistors are used for RF.



My dummy load is full of tab resistors clamped to aluminum heat sinks.

That's OK at low frequencies, but can have a high VSWR t VHF and up.


--
Anyone wanting to run for any political office in the US should have to
have a DD214, and a honorable discharge.

 

[Phil Allison]

"Cydrome Leader"

BTW:

I could have added a RC network ( ie a Zobel ) across the 1 ohm wirewound
to
bring it line at 3MHz and well beyond. A 47nF cap and any 1 ohm, 0.5W
film
type would do fine.

Same goes for the 100W 8 ohm power resistor, a Zobel using 68nF and 8.2
ohms
corrects it to beyond 5MHz if that was ever needed.

I wonder how much that would confuse the next repair tech?

Bizarro repairs are always fascinating to come across.

** I have sometimes added a Zobel ( ie 22nF and 56ohms, 0.5W) ) across the
output of an amplifier that did not need it.

Why ? As a tell tale in case of supersonic oscillation.

Incorrect installation and bad cabling practices can cause almost any power
amp to break into oscillation at an inaudible frequency resulting in rapid
overheating and failure of output transistors etc

It's nice for a repairer to have proof when this has happened.


..... Phil

 

[dave]

On 07/31/2014 05:52 PM, Phil Allison wrote:

"Cydrome Leader"

BTW:

I could have added a RC network ( ie a Zobel ) across the 1 ohm wirewound
to
bring it line at 3MHz and well beyond. A 47nF cap and any 1 ohm, 0.5W
film
type would do fine.

Same goes for the 100W 8 ohm power resistor, a Zobel using 68nF and 8.2
ohms
corrects it to beyond 5MHz if that was ever needed.

I wonder how much that would confuse the next repair tech?

Bizarro repairs are always fascinating to come across.

** I have sometimes added a Zobel ( ie 22nF and 56ohms, 0.5W) ) across the
output of an amplifier that did not need it.

Why ? As a tell tale in case of supersonic oscillation.

Incorrect installation and bad cabling practices can cause almost any power
amp to break into oscillation at an inaudible frequency resulting in rapid
overheating and failure of output transistors etc

It's nice for a repairer to have proof when this has happened.


.... Phil

I listen to distant AM radio in the shop. An amp that breaks into
oscillation (often caused by "replacement" actives with way too much
bandwidth, BTW) will cause interference on the AM (MW) radio. It is as
reliable an indicator of oscillation as fat blurry traces on the Oscope.
Also, the amp pulls more current than it should when it breaks into
oscillating. Now a SMPS without EMI suppression already makes the
interference so all bets are off.

 

[Phil Allison]

"dave"

I listen to distant AM radio in the shop. An amp that breaks into
oscillation (often caused by "replacement" actives with way too much
bandwidth, BTW) will cause interference on the AM (MW) radio. It is as
reliable an indicator of oscillation as fat blurry traces on the Oscope.

** Errrr - that is "parasitic" oscillation at radio frequency.

My post was about a quite different matter.

Maybe you have never seen it - it kills output stages in 30 seconds.




...... Phil

 

"There are no class D amp chips and the mode of operation of the output stage
is mysterious. "

Not really. That is only a one channel amp. Even operating in BTL, if you sit down and figure out a class D drive circuit using discrete components, the count is not all that high. Class DB might take a little ore but since this is a bass amp the filters can be miles away from the operating frequency so DB is not really all that much advantage. Did you stick a scope on it and find out just what the amp itself chops at ?

"** I have sometimes added a Zobel ( ie 22nF and 56ohms, 0.5W) ) across the
output of an amplifier that did not need it.

Why ? As a tell tale in case of supersonic oscillation. "

Good thinking if you suspect that. I take it you mean mainly in analog amps..

I have done similar in amps with fuses internal. When they do not have current limiting amd people stick like a 2 ohm load on them and shit, it ain't my responsibility if the waste more silicon and they pay. I almost never do that, but have in cases of multiple recalls. One multiple recall was due to a leaky muting transistor at the input of the power amp. I ALMOST got to accusing the customer but it was always the same channel. The boss said they shouldn't be turning it all the way up, but I disagree. They are entitled to use every last watt that amp can put out. Since it was after the volume and tone controls the problem would only show up at very high levels. The fucker was rectifying the audio and imposing DC on the output. But that was where everyythignweas under warranty, a rental outfit. there is no simply siying screw it and give them a refund to divorce a piece of junk. We would have had to actually replace it. Another company found that out about selling service contracts. If you pay your housde insurance and your house burns down, just refunding your premiums does not cut it.

>"Incorrect installation and bad cabling practices can cause almost any power
amp to break into oscillation at an inaudible frequency resulting in rapid
overheating and failure of output transistors etc "

This just happened to a buddy of mine recently. It was making a funky kind of distortion which cleared up at higher input levels. I said "Shut that off NOW !". Tunred out a speaker wire was running parallel to a line level input, but that was still downstrean of the gain control. It seems that luckily, the oscillation was high enough not to fry his tweeters. I cleaned that thing up and it is now quite happy, except for the one amp working into four ohms.

>"Maybe you have never seen it - it kills output stages in 30 seconds"

Think about what would happen with piezo or electrostatic tweeters. Even ribbon tweeters. When it happened to my buddy, it must have been high because if it was only 50 kHz or so, with the power, those tweeters should heve been toast.

 

[Phil Allison]

<jurb6006@gmail.com>

"Phil Allison"

"There are no class D amp chips and the mode of operation of the output
stage
is mysterious. "

Not really.

** Then kindly explain how the PWM signal is being generated.

Did you stick a scope on it and find out just what the amp itself chops at
?

** There is very little HF at the output to go on, just some harmonics at
300 to 500 kHz.

"** I have sometimes added a Zobel ( ie 22nF and 56ohms, 0.5W) ) across the
output of an amplifier that did not need it.

Why ? As a tell tale in case of supersonic oscillation. "

Good thinking if you suspect that. I take it you mean mainly in analog amps.

** Is there another kind ???

( Apologies to Jack Nicholson ..... )

"Incorrect installation and bad cabling practices can cause almost any
power

amp to break into oscillation at an inaudible frequency resulting in rapid
overheating and failure of output transistors etc "

>"Maybe you have never seen it - it kills output stages in 30 seconds"

Think about what would happen with piezo or electrostatic tweeters. Even
ribbon tweeters.

** Almost instant destruction.

When it happened to my buddy, it must have been high because if it was only
50 kHz or so, with the power, those tweeters should heve been toast.

** 50kHz or thereabouts oscillation does no harm to typical woofers - the
inductance of the VC protects them.




...... Phil

 

>"** Then kindly explain how the PWM signal is being generated. "

Relaxation oscillator to generate a sawtooth wave, comparator to chop it up into PWM. It's not that hard really. The chips that do it are just easier to use than to design it with discrete components. After tha you got your PWM square waves, then you design the drive circuit for the outputs. What's more, even though it is BTL, you only need one, it is simply inverted for the other phase.

Looking at the picture, there are plenty enough transistors and whatnot on the board to do it.

Of ocurse the question is why. Why didn't they just use some off the shelf chip for that ? To answer that we need to know things we are not going to find out. Amybe some parameter was too hard to get out of the chip, maybe their chopping frequency is lower than the chips like to operate at. Maybe certain effects were easier to implement, though I'm having a hard time fathoming any at the moment.

Or it may be attributable to audiophoolery. Some people like discrete components, myself included. I know damn well thaat what's in the chip is the same as what's in the transistor package, but...

Like if you bought a Marantz that uses an STKXXX instead of a Luxman that uses the same STKXXX, what did you get different ? Maybe us nuts just like to make the engineers work harder.

But really, among musicians you will find a healthy dose of audiophoolery, so it may be discrete for marketing purposes.

Of course in class D it doesn't mean a thing as far as I can tell. But then if ONEE golden ears can prove he hears the difference, the market will beat a path, to a certain door.

 

[dave]

On 08/01/2014 08:51 AM, Phil Allison wrote:

"dave"


I listen to distant AM radio in the shop. An amp that breaks into
oscillation (often caused by "replacement" actives with way too much
bandwidth, BTW) will cause interference on the AM (MW) radio. It is as
reliable an indicator of oscillation as fat blurry traces on the Oscope.


** Errrr - that is "parasitic" oscillation at radio frequency.

My post was about a quite different matter.

Maybe you have never seen it - it kills output stages in 30 seconds.




..... Phil

I know exactly what we are talking about. Radio Frequency starts at
10KHz. My traces fattens. The radio hisses. And the dim bulb glows
brighter. The dim bulb tells me to cut power way before any destruction
in 30 seconds happens. Modern IC opamps have too much bandwidth.

 

[Phil Allison]

"dave"

Phil Allison wrote:

I listen to distant AM radio in the shop. An amp that breaks into
oscillation (often caused by "replacement" actives with way too much
bandwidth, BTW) will cause interference on the AM (MW) radio. It is as
reliable an indicator of oscillation as fat blurry traces on the Oscope.


** Errrr - that is "parasitic" oscillation at radio frequency.

My post was about a quite different matter.

Maybe you have never seen it - it kills output stages in 30 seconds.



I know exactly what we are talking about.

** No you don't.

Radio Frequency starts at 10KHz.

** No it does not.

> My traces fattens. The radio hisses.

** That is RF parasitics - not FULL POWER supersonic oscillation due to
positive input / output feedback.

And the dim bulb glows brighter.

** ROTFL.....


..... Phil

 

[Michael A. Terrell]

dave wrote:

Modern IC opamps have too much bandwidth.

No, they don't. Some people just don't know how to use them
properly.


--
Anyone wanting to run for any political office in the US should have to
have a DD214, and a honorable discharge.

 

Fuck you.

You make an oscillator with perhaps two or three transistors. Then you get matchdf transistors and connect the emitters together. Whatwever their collectors pull current from becomes "on" time. IOr poisssiblky "off" time. I have coibsuidered boith actuyally becasue of trying to make my pet single ended push pull stage work, if you remember that. Diploma or not, I amn pretty sure it is beyond you.

How it works is a mystery my ass dude. What, you z tube jockey and fuse changer now ? I know better than that.

OK motherfucker - the circuit :

An astable mul;tivibrator switching at about 50 kHz. Generate a triangle wave by simple integration. REGUALTE IT, it determines the open anded gain of a class D circuit.

Apply the input to another leg of the OP AMP you built, or just mix it with the damn triangle wave. It simply does not matter.

One of the most impportant things about engineering is to understand what matters, and what does NOT.

You know, I ight just up and get a paper and draw up something right now and upload it so you can see it. I just want to illustrate (hmm, that might do iut wh ? ) that is just ain't all that hard. ;I can concieve the entire schematicv of it in my head right now with not trouble. Putting it into Spice is a different story.

Maybre I should make it apoint to learn Spice better, that way because you cna run siulations and whit, you would better understand some of my concepts.

 

Actually class D amps with less negartive feedback are excllent cnadidates for soft clipping. Just change the original sawtooth or triangle waveform. Flletten it at the peaks. What's more it is impedance independent. Anything that fucks with the gian of the output stae will change with load. When you do it this way, it is constant. ;

You can sett your watcxh to it.

 

[Phil Allison]

<jurb6006@gmail.com>

>"** Then kindly explain how the PWM signal is being generated. "

Relaxation oscillator to generate a sawtooth wave, comparator to chop it up
into PWM.

** Huh ??

You cannot fucking make guesses about what the schematic is.

Piss off fool.



...... Phil