ESP P101 using rod pcb very strange oscillation issue

brunogiovs

Member
2014-09-29 7:16 pm
Hello Guys.

I have a P101 rev c that is experiencing a very interesting problem.

When I connect only one channel from a source the P101 is works just fine (like just the right channel ). but when have both P101 connected to left and right a distorting gets added to the signal in both channels this issue happens even if I connect one channel and I connect just the signal gnd on the other channel.

Did you guys ever saw something like that? when the mosfets are bad can output current leaks into the input? any ideas? I would like to minimize the rework on my P101.

to give more context on what might had cause this issue , I had these amps running a load test close to the clipping using dummy loads and I had the ground earth disconnected. this might had cause an extra oscillation that could have stressed the amps. also after the test the dc offset was -6 volts on both amps.. and R19 blow.


Specs on the P101 during the load test.

My P101 is a high power pcb version C using Profusion mosfets ( ECX20 ).
it does not have c11 and c12 as profusion datasheet says that the have N & P mosfets at the same capacitance level. the amps were configured on mono blocks but sharing the same chassis. and only the chassis heatsinks included were connected to the earth gnd and both transformers have a internal shield that are connected to earth gnd too.

ps. now I'm testing the P101 using a lab grade power supply and I was able to reproduce the same issue when the amps is on the chassis using it's power suppy.

Thanks in advance.
 
Are you able to say what frequency range the oscillations are in (assuming for sure you have oscillations) ?

My limited observations of my own work says:

few hundred kHz - often a wiring issue, stray feedback from power supply or grounding or output wires to the input wires, output signal current getting on the ground of the input is a classic

MHz range - usually an issue with the negative feedback loop but unlikely in a completed design like this

> 10MHz and difficult to see at all - the MOSFETs have local parasitic oscillations
 

brunogiovs

Member
2014-09-29 7:16 pm
Are you able to say what frequency range the oscillations are in (assuming for sure you have oscillations) ?

My limited observations of my own work says:

few hundred kHz - often a wiring issue, stray feedback from power supply or grounding or output wires to the input wires, output signal current getting on the ground of the input is a classic

MHz range - usually an issue with the negative feedback loop but unlikely in a completed design like this

> 10MHz and difficult to see at all - the MOSFETs have local parasitic oscillations

Initially I see like a distortion on the positive side of the sine wave.
looks like a feedback... looks like something is getting in the input signal.
after it goes to more than 50 mhz or the signal due to the feedback becomes out of control..

you think it could be the mosfets?

Thanks
 

Ian Finch

Member
Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
.....it does not have c11 and c12 as profusion datasheet says that the have N & P mosfets at the same capacitance level....
ESP said:
....Although not shown, the transistors and MOSFETs are the same in this version as for the low power variant. The additional capacitors (C11 and C12) shown are to balance the gate capacitance. The P-Channel MOSFETs have significantly higher gate capacitance than their N-Channel counterparts, and the caps ensure that the two sides of the amp are roughly equal. Without these caps, the amp will almost always be unstable......
I think you may have misunderstood that you still need C11,12 except that presumably, they will be equal in value if the Mosfets have equal gate capacitance. Try it with the recommended values first and see whether oscillation persists. Project 101 - High Power, High Fidelity Lateral MOSFET power amplifier

The place to find the details is the ESP forum, because full construction info. is only available to people buying the PCBs. Not everyone here has done so.
 

Ian Finch

Member
Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
I should correct a few errors above: First, Exicon Mosfets do have different gate capacitances that you still need to correct for. EXC20 types have double dies and twice the capacitance of EXC10 types which are equivalent to renesas 2SK1058/J162. http://www.exicon.info/PDFs/ecf20n20.pdf

I wrongly suggested C11,12 were both used when only C11 was required for one pair or C12 for two pairs of Mosfets (or a double die device).

Fit C12 and see what happens to your instability.
 

brunogiovs

Member
2014-09-29 7:16 pm
I should correct a few errors above: First, Exicon Mosfets do have different gate capacitances that you still need to correct for. EXC20 types have double dies and twice the capacitance of EXC10 types which are equivalent to renesas 2SK1058/J162. http://www.exicon.info/PDFs/ecf20n20.pdf

I wrongly suggested C11,12 were both used when only C11 was required for one pair or C12 for two pairs of Mosfets (or a double die device).

Fit C12 and see what happens to your instability.

Hello , over the weekend I did replaced all transistors and driver, I also put c11 and c12.
it still oscillates when I connect both input sources . very interesting no?
If I connect only one channel to the amp works fine.. no oscillation..
 

Ian Finch

Member
Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
Have you posted this problem or read the P101 details at the ESP forum? It might be worth checking because you can be sure the amplifier works if assembled and wired up correctly. I'm not sure it was ever intended for, or tested with double die Mosfets. It should not be a problem but I have not checked it either.

Since oscillation only occurs with both amplifiers connected, I would suspect a common grounding issue. That is, you need to check the wiring layout of the speaker ground return leads. Each should only be connected to star ground separately and so should signal input ground. Errors here can cause hum but can also affect stability.

Make certain the power leads and ground for each amplifier are tightly twisted together and also the signal input pair.
 
Have you posted this problem or read the P101 details at the ESP forum?
ESP does customer support and gives extra private information to his customers. I see others coming here before going to their supplier. I can't understand why.
Make certain the power leads and ground for each amplifier are tightly twisted together and also the signal input pair.
Absolutely.
most of the problems stem from screwing up the grounding and most of those are due to "thinking" star ground.

Forget ground and use twisted pairs/triplets for ALL signal cable.

A cable that carries current will ALWAYS have a partner that carries that current back to Source. You must pair these up with LOW LOOP AREA.
 

brunogiovs

Member
2014-09-29 7:16 pm
Have you posted this problem or read the P101 details at the ESP forum? It might be worth checking because you can be sure the amplifier works if assembled and wired up correctly. I'm not sure it was ever intended for, or tested with double die Mosfets. It should not be a problem but I have not checked it either.

Since oscillation only occurs with both amplifiers connected, I would suspect a common grounding issue. That is, you need to check the wiring layout of the speaker ground return leads. Each should only be connected to star ground separately and so should signal input ground. Errors here can cause hum but can also affect stability.

Make certain the power leads and ground for each amplifier are tightly twisted together and also the signal input pair.

Sorry guys if you think I shouldn't be posting this questions here , I though this problem could be generic and someone could had solved.

Rod told me that he never saw this problem before.
 
ESP does customer support and gives extra private information to his customers. I see others coming here before going to their supplier. I can't understand why.Absolutely.
most of the problems stem from screwing up the grounding and most of those are due to "thinking" star ground.

Forget ground and use twisted pairs/triplets for ALL signal cable.

A cable that carries current will ALWAYS have a partner that carries that current back to Source. You must pair these up with LOW LOOP AREA.

Should I shield the input wire also? my signal cable is twisted but not shielded ?
 
A screened/shielded Coaxial cable is already close coupled and low loop area. But you MUST connect the screen/shield to the Signal Return at BOTH ends.
If you break the screen at an end, the current has to find another route that will be ENORMOUS LOOP AREA, or just won't work.
The current MUST RETURN to Source.

With a twisted pair one is less likely to break the return wire.
A screened/shielded twisted pair is better than unscreened. It's what they use for cheap microphone cable.
Screened star quad is better still and is used in good microphone cable. But it is roughly double the capacitance of other cables.

DF tells us that coax is better than twisted pair. I'm still not convinced even though he keeps reminding me I am wrong.
 
Last edited:

Ian Finch

Member
Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
Twisted pair leads are usually fine and self explanatory. I have seen some awfully long twisted pairs that wander everywhere in some commercial amplifiers yet noise is acceptable. I am not certain it is as effective as typical shielded wire at high RF frequencies, where Mosfets start to show their worst instability.

Perhaps it is worth trying shielded wire but I doubt the twisted pair is the cause. More likely, is the length and location of leads if they are otherwise correctly connected with short tails. A clear pic. of the insides of your amplifier would help spotting any layout/wiring errors, assuming you can post one.

Here is an interesting thread showing how not to build a P101 or any amplifier for that matter. I'm sure AndrewT will recall it and the advice given there: http://www.diyaudio.com/forums/solid-state/260374-grounding-issues-p101-rk27-wire-2.html