help reviving a Marantz AV9000 preamp

[Matt Syson]

Hi
Yes change the ceramics, starting with the ones on the AC side of the bridge. These can fail for no apparent reason. I would use 100 Volt or more types on any AC connections 50 Volt as original should be OK on the DC bits.
Interesting only 1 of the 2 fuses involved with this bit of the supply blows are you using T rated (antisurge) replacements? The 2 'halves' of the transformer secondary should be equal voltage, are they?
Matt S

[EricSan]

Hi Matt,

Thanks for your input. I checked again, both secondary legs of the transformer are putting out equal voltage 9.8vac on my meter. When the unit first died, both FB51 and FB52 fuses were blown. I replaced them both and the unit worked for 3 more weeks. Then they blew again and I replaced them, and it lasted for 3 days. After that, I replaced all of the electrolytic caps. The unit worked for 3 hours (something about 3's here...). Funny part is that after replacing the caps, only FB52 blew - FB51 remained intact. I am using slow-blow fuses rather than fast acting ones.

I then replaced all of the heat-sinked transistors (voltage regulators) and the rectifier (also heat sinked) that is connected to this set of secondaries from the transformer. It lasted another few hours before FB52 blew again (FB51 still good). There are a few transistors left that I have not replaced. I will replace all of the remaining transistors and replace all of the ceramic caps (0.01uF 50v) with some variety of film cap rated to 100v. Two more days for another DigiKey delivery...

I do think it is strange that only one side of the secondary was blowing a fuse...

Someone else mentioned that perhaps I should look for dead op amps. The only op amps I have found are on the subboard that handles digital-to-analog conversion.

Thanks,
Eric

[Matt Syson]

Hi
For now don't bother looking for 'faulty' op amps.
Start by measuring the current actually taken on the bit of the 'suspect' supply. Unfortunately you would need 2 ammeters but you could fit say half or 1 ohm resistors (wirewound to withstand initial surge) and measure volts across them. This will not be very accurate unless you have a RMS measuring meter but it would give an indication of how much is being taken.
Since the regulators look like being 1 Amp devices I would expect the maximum AC current to be a bit over 1 Amp (1.6 if the regulator is delivering rated current, which it shouldn't be).
For now you could pull the 0u1 caps out on the AC side to see if one of them is faulty. It may be an 'intermittant' thing so don't expect it to read a 'short' necessarily.
If left on for say 10 minutes or so, do any of the electro caps get hot? If so it may be in backwards.
Matt S

[EricSan]

Hi Matt,

Thanks for your continued input! You are correct, the regulators are all 1A (though they vary in voltage rating from 5v to 15v depending where in the circuit they are located). I am fairly certain that I was careful when mounting the new caps that I didn't get any of them backward, but its easy to check when I get home. I'll look there first.

Judging by the location of the 0u1 caps on the AC side of the bridge, am I correct in assuming that the power supply will function without them entirely? My guess is that they simply help to remove noise from the power supply and don't serve a "mission critical" role. Just looking for a "quick test" that I can do. If it won't cause any harm, I can just remove them and power it back up to see if the fuse lasts...

I don't suspect that anything will measure as a short since the preamp reliably works for several hours before the fuse goes. I would think that a short will kill the fuse promptly upon power-up.

[Matt Syson]

Hi
Yes the small caps on the AC side reduce the 'rf' switching noise from the diodes so are not essential to operation.
I feel your next test should be to measure the current actually taken by the 'naughty' supply section as I would not expect it to be anywhere near 1 Amp for a stereo preamp unit!
You could do a quick 'head count' of ICs powered by this part of the supply to get an idea of how much current it should take (assuming it has op amps!)
You need also to think where this extra current is GOING, as most parts would fry if taking more than about 2 watts without a heatsink. 10 Volts (or more) at 2 Amps must be making something hot (20 Watts).
I have known batches of 0u1 ceramic caps fail short or partial short even though they are 50 Volt working fed on 18 Volts DC.
Matt S

[EricSan]

Hi Matt,

Some additional background. The Marantz AV9000 is a 5.1 channel Dolby Digital, DTS, home theater preamp. Your suggestion of counting ICs and summing their current draw is a tall order as there are a great many ICs inside the chassis. Curiously, in each of the cases that the fuse blew, the preamp was processing the optical digital signal from my satellite dish. Just for kicks, I let it run all day yesterday without having any inputs connected and it was still running 15 hours later. If I activate the DAC circuitry by selecting that input and feeding it a signal, the fuse blows within 1-3 hours.

So, did a few more things last night. First, I pulled the fuse from the troubled side of the transformer secondary and placed it inline with a 0.382 ohm resistor (two 0.68 wire wound resistors in parallel from a previous project). Using my voltmeter to measure voltage drop across the resistors indicated a drop of anywhere from 0.62v to 0.71v. Dividing this by 0.382ohms reveals a current draw across the fuse in the neighborhood of 1.44 to 1.85A (on a 2A fuse). It seemed to vary each time I measured it, but I'm just using a $50 meter. After a short while, the fuse was noticeably warm to the touch. No wonder it keeps blowing.

I then removed the two 0u1 ceramic caps on the AC side of the bridge. This made no difference in the current draw across the fuse. Upon further inspection, I noticed that I had installed two of the small electrolytic caps backwards, so I turned them around. Still no difference. The larger of the electrolytic caps (13,000uF, 4700uF and 2200uF) were also warm to the touch on the top after running for a few hours. One small detail is that I was unable to find a 13000uF cap (C805 close to the rectifier on the schematic) so I replaced it with a 15000uF cap. Given the typical 20% tolerance on caps, I don't really see that this can cause a problem.

I then measured the current draw across the other side of this secondary winding off of the transformer (the "good" side). It also has a 2A fuse, but hooking things up the same way I did with the errant side revealed a voltage drop across the resistor of barely 0.01v. Current draw on this side is negligible.

I suppose I'll pull the rest of the 0u1 ceramic caps (there are about 25 of them) and replace them with higher quailty film caps and swap the remainder of the small signal transistors on the board.

Eric

[Matt Syson]

Hi
I commented last night that when using resistors to measure the current draw you should have the same value resistors for BOTH fuses fitted at the same time otherwise the 'real' fuse will actually be taking the bulk of the load. I don't know where the post disappeared to!!!
Is the supply to the optical input switched in any way such that only when selected does it get powered up?
An interesting challenge you have set here!
Good luck
Matt S

[EricSan]

OK, I had some time to do some additional measuring (though I still have more I want to try). I pulled both fuses and ran each in series with a 0.382 ohm resistor at the same time. Voltage drop across the "good" side was not measurable with my meter set on the 20VAC scale - it registered as 0.00v (set on the 2vac scale, the third decimal just seems to fluctuate on a random basis). The "bad" side measures 0.5vac (1.3A draw) with the pre-amp at idle and with no input.

When I switch the input on the preamp to one that required digital decoding, the current draw on the good side remained unmeasurable and measured 0.65v (1.7A draw) on the bad side.

I then fed it an actual digital signal to decode and current on the bad side jumped to 2.15A - no wonder the 2A fuse doesn't last long. Current draw on the "good" side remained at nearly zero throughout this entire test.

I also measured the voltage output from the regulated power supply at two cable harnesses on the left side the schematic posted above. Each pin on the connector measured spot-on for its voltage as specified in the schematic. Because there are so many pins, I am unable to measure current draw across each pin, but the voltages hit the specs exactly.

What I didn't get to try last night was to measure current draw with and without the digital decoding sub-board connected at all. There is one power harness that provides power to this sub-board. I want to pull this harness and see if that makes any difference in the power draw on the power supply board.

At this point, I am still suspicious of either the caps or the remaining transistors in the power supply. I just need some time to cross reference all of the part numbers and order them...

[Matt Syson]

Hi
FB51 and 52 (if I get the numbers right, both 2 Amp fuses) should have practically the same current through them. If one is zero, from what I think you say, then that is the problem. Either the transformer half secondaryu is not goving the correct volts or one of the diodes in the bridge is open circuit, or a ripped track??
Matt S

[EricSan]

This is the part I can't figure either: The two halves of the secondary winding are both producing 10vac (matches the specs), but the current draw is not equal across the halves. I have replaced the diode bridge that this secondary winding feeds and the filter caps. With all of my work dismounting and mounting parts, I did lift the pad off of the circuit board in one or two places, but when this happened, I jumpered that trace with some wire to make sure the electrical connection was not broken. It would have been nicer if they used plate-through holes on the PCB... I have looked for other areas where a trace would have been disrupted and not found any.

Though I did replace the rectifier, I have not measured the DC voltages coming out of the rectifier. I only know that the AC voltages going into the rectifier are on target.

I suppose this may reach a point where the pre-amp ends up being replaced, but thought it would be a good learning experience to try and track this one down. It was used when I purchased it about 5 years ago, so my economic investment is not large, but it just seems a shame to give up on it.

Eric

[Matt Syson]

Hi
I think you are getting close, don't give up now!
I would expect the 2 fuses to be contributing within say 10 percent of each other. I would suspect a cracked solder joint or a bit of track you have missed.

Good luck
Matt S

[EricSan]

I think I've found where that 2A of current is going. Last night, I measured the temperature of the heat sinks on the regulated power supply (voltage regulators are mounted to the heat sinks). The results mirrored the difference in current draw across the 2 fuses.

Q804: 36c
Q806: 45c
Q808: 48c
Q809: 46c
Q805: 54c
Q807: 49c

I tend to think that below 45c is a more normal operating temperature. The 54c definitely stands out to me. I checked all of my solder joints and everything looks good. I carefully re-melted several of them just to make sure that the solder flowed properly and spent some time looking for sloppy solder that may have shorted across the traces on the board, but didn't find anything.

I'll order caps and the remainder of the transistors today if I can find time.

[Matt Syson]

Hi
If you open the link which is shown as D808 then the current through the 2 fuses will fall to a pretty low value.
You must be able to measure the same current through both fuse FB51 and 52 otherwise the current will be too great as it tries to power the whole circuit. The relay in the -14 Volt feed I presume is operated when you have an extra bit of decoding going on which will increase the load on this part of the supply. The fact that other bits run quite warm is possibly intended, consumer gear is all about cost so allowing parts to run hot rather than a bigger heatsink is commonplace.
Get the current through the fuses equal and you have probably cracked it. Either duff diode, broken track or the winding of the transformer is 'lazy'.
You say it all works in which case replacing transistors is a waste of time, the problem is more basic.
Matt S

[EricSan]

Hi Matt,

Still looking to find a dead cap. I just ordered a pile of 0u1 100v film caps and will replace the ceramic ones when the order arrives. I'll be extra careful removing the old ones and see when happens with power-up and current draw.

I'm not sure what you mean when you talk about opening the D808 link. What would you expect that I learn by opening this link and observing the effects other than seeing the draw across the fuses drop.

The relays in the power supply don't open and close with the input selector, they all operate upon power-up and stay closed until power-off. Or at least I presume they are operating correctly as I haven't specifically tested each one.

The part that I find most curious is that the power supply is providing it spec'd power, yet still drawing too much current. If the new caps don't do the trick, I'm not sure what else to look for.

[Matt Syson]

Hi
You MUST find out WHY the 2 fuses are not carrying approximately equal currents, this is the main reason that one fuse blows as the 2 should be sharing. You need 1.6 TIMES as much AC current to produce the DC Thus for 1 Amp DC you need 1.6 Amps AC, which is where the fuse is located.
Either the measuring is suspicious (as you have only 1 meter you need to do the resistor thing on BOTH fusefolders at the same time) or there is a relatively simple fault in wiring, rectifier diodes or the transformer winding..
Concentrate on this part otherwise you will never get the fuse to hold in at 2 Amps.
Forget the caps for now as I expect they are actually OK.
Matt S

[EricSan]

I will spend some more time with the circuit board tonight. I do have one extra rectifier - I always order more parts than I actually need because I've been shorted in the past.

I don't know how to tell if the transformer is having a problem since all I can really do is to measure the voltage on each winding (which do measure as specified). I do have two meters and have been able to measure the current across both fuses at the same time. Nothing I have done so far (disconnecting the digital sub board, changing inputs, etc) has brought me any closer to balancing the current draw across the two halves of the transformer secondary winding.

I will also double check the board with the transformer and fuses mounted to it. Most of my efforts lately have focused on the regulated side of the power supply.

[Matt Syson]

Hi
The secondaries are about 10 Volts so you could put a 12 Volt car bulb, say 10 Watt rating or even 21 Watt on each half at a time and measure volts (about 10) and the current AC. You are looking for equality rather than an absolute level.
Do this without either fuse fitted. This should reveal both windings giving same current and voltage hopefully.
The only other 'stuff' involved will then be the rectifier, the fuse holders and the wiring, either wire or circuit board traces.
Matt S

[EricSan]

If removing possible problems constitutes forward progress, then I made some progress last night...

I pulled out the transformer and played with it all by itself. Both legs on the secondary winding of the transformer are producing 10vac, and both legs on the secondary winding are able to power a 7w bulb and a 27w bulb (a 12v dual filament automotive light bulb) to equal brightness. Figuring that the 27w filament represents a current draw of about 2A, it seem that the transformer can deliver the goods as spec'd. Cross that one off the list of potential problems...

I then pulled the regulated power supply and soldered a few leads to the DC pins of the rectifier. With no fuses installed, no resistors in the path of the secondary windings, the rectifier is producing -12.47v and 9.95v. Specs indicate that this should be 10.8v and -14v. Seems pretty close to me - cross another probable cause off the list...

Now, the interesting part is that we are both assuming that both sides of the secondary *should* be drawing equal current at all times. I don't know this for sure, it just makes sense given what we know...

I then placed the 0.366 resistor in path of the secondary windings (both of them at the same time, still no fuses) and the rectifier voltage then measured -12.43v and 8.47v. Voltage drop across each of the resistors measured 0.0v and 0.32v (0.9A) with the preamp in stereo mode and 0.44v (1.20A) when digital input is selected and no signal is present. This is much less current draw than I measured last week with the preamp at idle...

I then provided a digital signal to decode, still no fuses, both resistors in place, digital input selected and being fed an optial digital signal. Voltage drop across the resistors then measured 0.0v and 0.47v (1.28A). At the same time, voltage from the rectifier measured -12.0v and 7.0v.

I am not sure what is going on as the idle current I measured a few days ago was closer to 1.3A and when actually decoding a digital signal was topping out at 2.15A. I didn't measure anything like that this time. I wonder if perhaps there was a cold solder joint that I happened to correct last night while attaching wires so that I could measure.

So, the two sides of the secondary are still not balanced (I don't know for a fact that they should be), but the "bad" side is not drawing anywhere near the current that it had been drawing previously.

I will replace the ceramic caps tonight and see if that makes any difference and if the the current is still sufficiently below 2A, I'll just close it up and see how long the fuse lasts...

I have inspected the board and haven't found any obvious solder blobs that shorted two pads together or any obviously broken traces.

[Matt Syson]

Hi
Interesting and still puzzling!
The rectifier action is such that for the positive you have a transformer with centre tap and are using 2 diodes to get a full wave rectified supply.
The negative rail is the same but the diodes are obviously reversed.
You do not get the same DC out as the two Dc rails are pulling different currents, hence the 13,000 uF on one side and a small on the other.
I still don't get why one fuse is passing no current. What happens if you remove the other fuse? Do you get a low rail, nothing or is there something else going on?
Matt S

[EricSan]

I, too, was puzzled by the difference in size for the caps after the rectifier. One is 2,200uF and the other is 13,000uF. I remember being surprised that I only need one "big" one and that the other one was so tiny in comparison. For the other rectifier in the circuit (which behaves itself well), the caps that follow are equal in size to one another (4,700uF).

I don't have much previous experience in this area, but I have built a DIY class A amplifier Pass Labs a40 amplifier parts & vendor list and the caps following the rectifier are always equal and both legs of the secondary power supply draw significant current, but that is to be expected.

Perhaps using the values of the caps as a guide (2,200uF and 13,000uF), the two sides of the secondary in the preamp are not meant to draw anywhere near equal current after all...

I can check again tonight, but my recollection is that if either of the two fuses from the 10v winding on the transformer are removed, the protection circuit (or what I am assuming to be the protection circuit) activates, the relays open, and the preamp shuts down. Perhaps this "other" side that doesn't draw any current is meant simply to power the protection circuit, thus when things are operating properly, there is no current draw across this side. A few days ago, I disconnect all of the jumper cables leaving the regulated power supply (left side of schematic) so that I could measure the output of the power supply. Without these cables in place, the relays don't engage and the unit doesn't power up. I suppose that is what the POWER and PDOWN pins are for. When the appropriate voltage does not come back to the regulated power supply, the relays open and it goes dark...

[Matt Syson]

Hi
You have to think of this section as being a positive rail with a heavy load and a negative rail with a light load, cunningly using the SAME transformer windings. Yes it looks like a typical 'split rail' supply but it is intended to be uses with dissimilar loads. Not magical at all you could argue that the windings when supplying the negative low current rail are actually 'over specified'.
Are you sure you are measuring nothing on one of the fuses, there isn't a link underneath or anything?
Check out the 'full wave rectification using 2 diodes and 'split' windings' to get your head around what is bhappening. Then do the same for the negative rail then you will see that it 'looks' like you have a bridge rectifier.
Note the small diodes used to power the 'mains off' mute circuit, the two separate diodes below the bridge.
Matt S

[mixedupsteve]

Eric,
Did you ever replace the caps that you had backwards? I had a similar situation where I installed a cap backwards which causedthe fuse to blow like you would expect, I reversed it and still had fuses blow from time to time. It was a higher voltage tube power supply but installing the cap backwards did ruin it.
Just curious.

[EricSan]

Hey Steve,

Thanks for your suggestion. When I noticed the reversed caps, I did remove them, test then, and then put them back in. Both of them did measure correct capacitance, but I didn't mark which ones they were... So, this is not something that I can go back to check easily because I don't remember which ones they were. I have a vague recollection that they were both 10uf 50v caps, but there are a bunch of those... Tube power supplies can get pretty serious, but I'm hoping that there wasn't any real damage to these low voltage caps.

Matt: I see what you are saying about the rectifiers... I've probably been incorrectly using some of the specific terminology here. The more I think about this and look at the schematic, the more I become convinced that the center-tapped 10v windings are intended to provide different levels of current for different purposes. I'll double check the fuse that appears to be carrying no current.. With my meter set to the 20v range, it measures 0.00v. If I set it to the 2v range, it measures 0.00x and x seems to vary like a random number. Seems that we are the design limits for my $50 meter.

With the "improved" behavior and lower current draw that I measure last night, I'm beginning to think that I am actually finished here and that the preamp will continue to function. The maximum draw I was able to create was a pinch below 1.3A, something that wouldn't seem to create problems for a 2A fuse. Maybe I'll try to find a 2.5A fuse for a little extra headroom, but that won't actually cause a problem if greater current does actually flow.

Since I have the new coupling caps, I'll probably put them in tonight and see if anything changes...

Eric

[Matt Syson]

Hi
If one of the two fuses is not passing any current as you say, then logically either C801 or C802 (0,01 uF caps) will not have any voltage across them. They should both have 10 Volts AC on them, measured to ground of course.
If they both have 10 Volts AC then one part of the bridge is not getting this 10 Volts so if 10 Volts is appearing on BOTH 'AC' terminals on the bridge then it is faulty.
I presume the things marked WB (xxx) on the schematic are wire bridge or wire linking between boards of some sort. Conductor 1 or 3 on this link must be 'open' or some part of the connection from the fuseholder to the bridge rectifier.
You could simply fit a 2.5 Amp fuse, but that is a simple 'cop out' and you have not really got to the bottom of this which should be relatively simple problem.
Check out the resistance from one AC terminal on the bridge to it's relevent fuse, then the other, do you get close to zero ohms?
I hope it is not a stupid question, but are you sure the bridge rectifier is in the correct way around, or if it was the fuse blowing earlier may have opened up one or two of the diodes in the bridge perhaps?
Matt S

[EricSan]

Hi Matt,

I'll measure C801 and C802 tonight. Currently, these caps are not in the circuit as I yanked them a few nights ago looking for a quick fix. Its easy enough to make this measurement. I expect to find the appropriate voltages there as the rectifier seems to be providing the correct voltage levels.

Yes, things marked WBxxx are multi-wire harnesses that run between physical boards. Each physical board in the preamp is indicated by the heavy dashed line in the schematic. Thus, PB54 is a small circuit board that contains the transformer and several fuses. A wire harness then connects the secondary side of the transformer to the regulated power supply main board that is designated P804 PRIMARY BLOCK. PB74 is a small PCB that is mounted to the power swich on the front of the chassis. Three of the voltage regulators are mounted on their own small PCB and marked P874, P884, and P894. Many of the voltage measurements I have made have been made from the backside of these wire harnesses, so there doesn't appear to be any breaks in the continuity of these wires. I will double check these tonight. The key, as you indicate, is to make sure that conductors 1 and 3 are each carrying 10Vac.

I am certain that the rectifier is oriented properly. The outer casing is marked + ~ ~ - in the same order as the one I extracted. The rectifier is receiving 10Vac and producing 10v and -14v as indicated in the schematic, do I don't think there is anything wrong with it.

I know that a 2.5A fuse is a cop out. Part of me is getting frustrated chasing down something that I'm no longer certain is a problem. When fuse FB52 was carrying 2.1A - that was a problem. Not sure what I changed, but now it seems to be carrying a maximum load of 1.2A.

I'll make a few more measurements tonight and see what I find.

[EricSan]

Ooops - I just realized that while making the measurements of current across the fuses that got me so excited last night were made with my meter in DC mode, not AC mode. Measurements of voltage drop made in AC mode are much higher than those made with the meter in DC mode. So much for making progress...

[EricSan]

After being nearly ready to give up, I found the problem!

Matt, it was your suggestion to check for continuity between the AC leg of the rectifier and tap from the transformer. I found one joint on the PCB mounted to the transformer that had a hairline crack in it. The trace and solder blob was disconnected from the post for the wire harness. I must have broken it loose during one of the many times I took it all apart. I re-melted the solder joint and then re-measured the current across the fuse on the "bad side." This time it measured 0.26Vac and when I gave it a digital signal to process, it went to 0.34Vac. This keeps current draw through the fuse to below 1A. I'm back up and running now!

Funny, the preamp works just fine without the FB51 fuse installed at all. I suppose that it is for the protection circuit. When I first opened the chassis back in August, both FB51 and FB52 were dead. There were several dead electrolytic caps in the power supply, I suppose when these failed, it caused the protection circuit to activate and popped both fuses. Then, the intermittent solder joint was causing an ongoing problem with FB52. After fixing this broken solder joint, the voltage coming off of the rectifier held more steady at 10v and -12v. Earlier, the positive voltage was dipping to 7v.

Matt, thank you for your time and encouragement! Without your help, I would not have been as successful. Steve, thanks also for your suggestion to double check the caps.

I was just about ready to fork over a few hundred dollars to pick up another used preamp... Thanks again!

[Matt Syson]

Hi
I am glad you have got there, I hate to leave 'mysteries'!
It probably is fine with one fuse missing, until it overloads the other because it is not sharing. The voltage 'pre' the regulators was sufficient with only half wave rectification.
You would have found that if your mains is at the 'bottom' of the specified tolerance for this unit that it would hum pretty badly although with 15,000 uF it may not have been so bad.
Now for your next project!
Matt S

[EricSan]

Ha- I didn't have it up and running enough to notice any hum withthe power supply was running on only half-wave rectification. I was measuring the resistance of the path between the transformer and the rectifier and one leg was a solid connection with only 0.004ohms resistance. The other leg sometimes measured 0 ohms, sometimes measure 190 ohms, and sometimes measured 80 ohms. I spent some time measuring different segments of the connection and finally narrowed it to the board with the transformer. That's when I noticed the crack in the solder joint. Its a place I never really looked at before because I had never fiddled with that part of the board.

The next project is a long and on-going one. I am building a 5 channel audio system for my basement theater. Five Pass Labs Class A 100w monoblocks and five floor standing 5-foot tallk dipole speakers. I have all of the speaker drivers, but not all of the cross over parts. I also have most of the parts for the amps except for the heatsinks and chassis components. All of the output Mosfets are precision matched and labelled just waiting to be mounted to giant heatsinks. The nice thing about Class A amps is that they contain relatively few parts, so when something goes wrong, the trouble is relatively easy to find since there are so few places to look. The pre-amp was somewhat overwhelming due to the large number of circuit boards and the extreme quantity of parts of each board, plus my lack of familiarity with the design concept and philosophy.

Thank you again for your very kind and generous help! I appreciate it very much!

Eric