Samson Xp308i problems

[diydidi]

Hi.
I have this unit for repair. Repaired the smps, but there is still something fishy going on. When probing the output, no load, the limiter kicks in as it should. I can see the amp clipping and then the limiter reducing the clipping to a nice sine wave.
However, connecting an 8ohm load or 8ohm speaker, the output is pure until the limiter starts kicking in. The limiter doesn't produce a clean signal at the onset of clipping, but instead the output bursts on and off, as if the power supply isn't able to deliver the current. Measured the sec voltages when this happens, and they drop to 17 V instead of the constant 26V as indicated by the schematic.
The fan also cycles on and off when this happens, and the power supply gets quite hot.
I have added the schematic. Any help??

[soundguruman]

There's a switching power supply, the schematic has the output voltages only.
One strategy is to pull the power supply components and test / substitute.
at least some of them can be tested
Another is to use freeze spray on the components one at a time.
There is probably the same part failing in many of those units...
When it's cold it fails, when it's a little warm it works, and when it's hot it fails again.

While it "does" work, I would hit this and that power supply component w/ freeze spray.
This is one way of finding a thermally intermittent component. (Or heating it)

But then again it may be a bad solder connection also.

If you bang on the circuit board does it cut in/out?
This can be an indication of a bad solder connection. The connection may work / fail depending on temperature.
You can also hit solder connections w/ freeze spray.

All in all an intermittent is the difficult type of problem, or thermal intermittent.

But you could suspect that hot parts/ connections are the most likely to be intermittently failing as a result of temperature.




Another bitch about it is the power supply transformers
You do not have waveforms on the schematic.

And parts can test good, then fail when under power / temperature.
This makes it especially fun.

[km6xz]

How are you testing the power, with a scope? Are you grounding the scope shield? Are you using sinewaves? This is a Class D amplifier that has differential output. Neither line can be grounded. You can test using two probes, disconnecting the ground/shield clips from the probes then set the scope vertical mode to SUM or ADD and invert one channel of it has a channel inversion switch. Set both vertical attenuators to the same voltage, say 20v/div.
Another problem is power is not measured with a continuous sinewave like an A/B amplifier. Power is measured by a short burst, usually 4 cycles. A Arb generator or function generator is normally used for testing Class D amps and distortion measurements are not done at higher power because there is not enough time for auto-notch filters to tune out the fundamental so either a fast spectrum analyzer is used or a special fixed tuned notch filter is used with an harmonic distortion analyzer. If it has low distortion at 1 volt peak to peak output, it will have low distortion at 20 volts with a switching amp. The reason being the same linearity is being exhibited at both levels, with the same peak pulse voltage, the difference between high and low power output is pulse width. What does is sound like when driving a normal load without any test instruments connected. I suspect it is working properly. Try variable amplitude music up to onset of clipping.

Testing and analyzing Digital amps takes a bit of different approach and test procedures, as do the measurement standards which are more geared to the old peak music power rating, which is a better indicator of the effective loudness for digital amps. Grounding one side of the amp output with a scope lead is shorting one side to ground so the protection circuit must be pretty good to survive full power into a direct short to ground.

[diydidi]

I'm aware of all of the grounding issues and Class d. I use a differential probe for these measurements. The amp works great if I connect a 100ohm resistor in series with the 8ohm speaker. The limiting works great too. The darn thing plays, just the limiter doesn't work with only the speaker connected.
Checked all components. Could a faulty transformer be the culprit? It too gets hot. I have one of those Dick smith LOPT testers. I 'ring' the primary winding and get a low ring. I did this test with transformeer removed. I think this transformer is the problem
I know it is usually highly unlikely with smps, but possible with this one....

[km6xz]

If it can't supply the load, and rails sage a lot, it is the power supply. It is rare, as you say, that the transformer fails unless physically damaged(cracked ceramic metal core). If it is really getting hotter than normal, losses of efficiency in the secondary is a good place to start. The PFC chip will shut down the primary when over voltage occurs via Q3. Some of the things to look for when noticing heat on the tranformer are the secondary filter caps(1000/50v) or the diodes D13/14. If you want to isolate the problem to the powers supply, pull the down-leg connection to L3 and L4 and attach a dummy load that can handle the power, about 12 ohms to ground. If a current problem it will clamp on that, or stay up if I am wrong and it is in the power amp. As a matter of normal repair, those secondary diodes should be changed any time you repair a SMPS. If they are stressed the junction, over time changes in off capacitance and there is a gradual increase leakage current. Changing the junction capacitance on a high frequency power signal drops efficiency a lot and loads the transformer, thus heat. Excess heat can crack the core.

[J M Fahey]

That PSU *must* be able to supply some 5A into the amps at rated power (100+100W) so load it end to end (+28 to -28) with a 10 ohm 200W`resistor and check whether it can do its job or not.
Or to avoid reaching the limits, use a 20 ohm 100W one, whch *should* cause no problem at all.
Kludge the resistor with some spiral wound nichrome wire out of some heating appliance and submerge it in water inside a Pyrex glass container or something.
Even some enamelled iron pot will do, but of course nothing conductive.
It might even be a plastic bucket if you dare.

EDIT: I'm squinting my eyes in this small netbook screen but can't find the limiter circuit, where is it?

[paleeman]

Hi diydidi
As usual, Samson came up with another unnecessarily complicated PSU.
I would inspect the first section, the PFC. It should deliver approximately 400V at C11 (120uF/450V) at all times. If it is steady at rated load, then you can skip to the next step. The power supply is regulated from the secondary side with an optocoupler marked PC3. The current limiting is monitored only inside U3, so BMHO either the U3 is faulty or some of the external components. The Q4 should deliver 15V to U3. Although this voltage should not make a difference to the output if it stays within operating range 9 - 23V.

Does the PS get hot in idle state?

[diydidi]

Hi diydidi
As usual, Samson came up with another unnecessarily complicated PSU.
I would inspect the first section, the PFC. It should deliver approximately 400V at C11 (120uF/450V) at all times. If it is steady at rated load, then you can skip to the next step. The power supply is regulated from the secondary side with an optocoupler marked PC3. The current limiting is monitored only inside U3, so BMHO either the U3 is faulty or some of the external components. The Q4 should deliver 15V to U3. Although this voltage should not make a difference to the output if it stays within operating range 9 - 23V.

Does the PS get hot in idle state?

Hi Paleeman.
I checked opto coupler. Tests fine. Interesting thing is, ZD1 keeps blowing when I drive the amp into clipping with 8 ohm speaker.
PS doesn't get hot in idle state. Q4 tests fine too. Maybe I should replace U3 and see...
I'm just skeptic of the Transformer. I need to get another good unit like this and A/B the transformers with my ring tester.

[paleeman]

If the ZD1 keeps shorting, it means that the auxiliary coil on the primary is delivering higher voltage than 27V. This can be caused by a faulty diode D17, or higher than normal (around 50kHz) oscillating frequency of the PSU. I would put my money on U3, definitely.

There is another possible scenario. If the windings are damaged on the switch mode transformer, as you suspect, it can result in U3 rising the oscillating frequency, trying to compensate the voltage drop on the secondary side, but unfortunately, it rises also on the primary side. I reckon, this is the reason for ZD1 shorting.

What I would try next is to find a point on the PCB where you can disconnect the power supply from the amplifier section, and test the PSU with a dummy load (a 100W light bulb will suffice ). This way, you can find out if the PSU is OK or not.

[paleeman]

Ooops. Sorry for the dummy load thing. I have just noticed that many people before me suggested this.

[ips]

Hi guys. I know it's being a while since last post but have anybody figured out the transformer or substitute? I have few broken units all need transformers.

[J M Fahey]

How do you know the transformers are broken?

It's very unusual.

[ips]

I swapped them between good and bad unit. Problem follows bad transformer. I think transformer getting cooked because of some other failure.

[J M Fahey]

But these use SMPS, not plain "transformers".

Did you exchange the actual ferrite cores only, or the entire SMPS power supply modules?

SMPS can and do fail, but a whole batch of bad ferrite cores is inusual (never heard about such thing).

[ips]

This is T2 Transformer(ERL-3541B7 6000WH±10%) P/N 411-XP308I-744 or 411-XP308I-833