Ampeg SVT 4 Pro 10V dc offset

[dwmorrin]

I have a Ampeg SVT 4 Pro here with 10Vdc at the output of both power amps.
Without load or signal:
+/-V measures +/-85.
+/-15 measures +/-15.
dc test points in the power amp page 1/3 look good until the "A_HI"/""A_LO" and "B_HI"/"B_LO" (inputs to the gates) nodes, which are all about +10V.
All of the P chan devices measure +9.97V on the gates, and all of the N chan devices measure +10.13V on the gates. (both amps are same)
All of the output devices sources are at 10.06V.

According to the schematic, the "HI" node should be at +3.5V, and the "LO" node should be at -3.5V.

Q102's voltages are:
E: +84
B: +83
C: +10.13
Q105's voltages are:
E: -84
B: -83
C: +9.97

Q103/104 have +/-5V on their respective bases (test point 6/7), and the dc levels appear normal before that.

I have been combing through the board powered down looking for shorts, opens, or anything unusual, but I can't find anything. In-circuit diode tests on the transistors look normal.

Where should I look next?

[hodgy]

How do the output mosfets and the 47 ohm resistors on the Gates look?

[nevetslab]

10V DC Offset

With Q105's collector at 10V, I'd partly suspect Q105. But, as you're getting similar Vbe readings on Q102 & Q105 (I'd presume those to be more like 0.6V Vbe readings, not 1V), it's behaving like there's a short across the Gate Drive buses.

Under normal conditions, where the HI node is 3.5V and the LO node is -3.5V, that shows the bias circuit network Q106/Q107 and it's host of supporting resistors and bias trim pot are working. In your case, these aren't allowed to turn on and spread the potential of the gate drive busses. You do have the protection circuit that shuts down the gate drive via Q108, Q109. It's almost like Q108 is turned on hard, pulling the output rail to the HI node. But, to get to 10V, something else is turned off. I'd look at components in around Q106, Q107 for starters, and also around the voltage gain stage ahead of this...Q105, Q104, Q104, Q103, Q102.

From the HI & LO gate drive rails, you do have back-to-back 10V zener diodes D114, D115, D116 & D117. But, your gate drive rails are effectively shorted, so just a coincidental value probably having nothing to do with it.

Surprising you have this on BOTH channels, and not just ONE channel. Now, looking back at my schematic, IC1 IS COMMON to both channels...an NE5532 Dual Op Amp. The other op amp circuits in the front end are unique to each channel, but NOT IC1. Suspect?

Is the amp in Bridge Mode? Do you get the same 10V offset in both Stereo and Bridge Mono, looking at the HI and LO gate drive rails?

Is the amp pulling current...higher than nominal off the Mains?

While you have the main amp PCB assembly out of the chassis (I presume), and this has nothing to do with your DC Offset problem, I'd have a look at the solder joints of the main Bridge Rectifier, and the solder joints on all the rear panel phone jacks. I repair these often, and I regularly find fractures at these locations. Also, the front panel handle hardware is just awful. I replace their # 10-24 PHMS with Socket Cap screws, flat & split lock washers, with a little dab of loctite before driving the screws back home into the handles, so they never come loose again. Just normal routine maintenance items on Ampeg SVT4 Pro amps.

[dwmorrin]

Yes, the global aspect of the issue is making me think it's something unusual - like a broken trace, or something like that. IC1's output is about 0V, so I wasn't suspecting that... but I won't rule it out.
So, I decided to stop using the chassis ground for my 0V reference, and things are looking different!
If I take the voltage between A_HI and A_LO (red lead HI, black lead LO) I get +6.75. Looking at the same nodes using a "ground" point over in the power supply section, I'm still getting +10V on both... so that might explain the global aspect.

Fuse hasn't ever popped. Current draw appears to be normal. Haven't measured, but passes dim bulb test fine. The amp is in the "stereo" mode for all of these tests.
Hodgy: Output MOSFETs and their gate and source resistors test fine in-circuit. I haven't felt the need yet to pull anything.

I'm gonna scratch my head over the voltage readings for awhile. Any more suggestions welcome.

[Jazz P Bass]

The voltage readings 'should' all be the same no matter where your probe is connected.
The schematic shows all three (chassis, signal, circuit) grounds tied commonly at the transformer center tap..
I guess how they get to that point may be the issue.

[dwmorrin]

Well, the problem appears to have been solved.
I assumed that I could leave the other pcbs disconnected while troubleshooting the power amp. After reconnecting the other pcbs, the voltages look normal, and in fact the amp seems to be working OK.
The amp came in with "weak output" as the reported problem, and I did verify that it indeed had weak output before opening it up. Preamp tested OK, power amp had barely any voltage coming out of it.
Now, the output of each amp looks good with 4Ω load, so while that's good news, I do hate it when amps seem to "fix themselves" in the event it's really some intermittent problem.
I'll double check those fx loop jacks, as always.

[nevetslab]

The fact that you're getting 6.75VDC between the Gate Drive busses indicates the bias circuit is working, and there's no protection clamp diodes turned on.

Take a look at the Vce & Vbe readings of Q102 & Q105 at the top and bottom of the Voltage Gain Stage. They should be very close to the same....measuring just Base & Emitter, Collector & Emitter, without reference to your circuit ground. Also look at the voltage across their emitter resistors...47 ohm. That will tell you what the current is in the string. You should be able to also compute the current thru the two emitter resistors in the center of the string, where IC1 feeds the string (R120 & R121).

As for a ground reference, there are several around IC1....bottom of R106, also junction of R114, R115, R116 & R117 around IC101D. Also bottom of R102, by IC101A.

If from any of those points, whatever is convenient to hang your Black lead of your DMM onto, you should be getting near 0V potential at the output of the amp. I'd stay on the circuit side, where all you 0.47 ohm 5W Source resistors common up, rather than looking at the output connector of the amp.

The fact that you're getting this odd 10V reading using Chassis as your ground does suggest a continuity issue. Open trace, fractured solder joints (always suspect. Another ground reference is where R131 & R132 common up at the protection circuit...6.81k resistors, each in series with a signal diode off the bases of the NPN/PNP short circuit protection Xstrs QQ108/Q109. I'm guess from these ground references for the amp circuit, your output is pretty close to 0V. I forget what the typical DC offset is on the amp.....+/- 10mV? +/- 50mV?

On Page 3 of the Power Amp PCB Schematics, following the CenterTap lead of the Power Transformer, Connection J20 shows the three Ground Reference symbols being common (Signal, Circuit & Chassis Grounds). From the power supply Ground, where the C/T lead join with the power supply caps C4, C5, C6, C7, C8, C9, etc, there shouldn't be any voltage difference from there to the local circuit grounds around IC1 & IC101. So, I'd think you're looking to find where this 10V potential is coming from.

Hope you marked all the primary leads to the main PCB that you have to disconnect in order to extract that main amplifier PCB assembly from the chassis. Some of the amps only have domestic primary, others the full universal primary, with more leads having to get to the correct PCB terminals.

From your Power Supply Ground reference, and assuming your output (circuit side of output relay) is at/near 0V, do you measure the same at the output jacks? I assume the relays pull in. They won't pull in if there's significant DC voltage present on the output of the amp. I haven't calculated what the voltage threshold is for inhibiting relay closure, but I'm sure it wouldn't allow 10VDC on the output.

[nevetslab]

Weak outputs.....had the same reports from our rental units. Always ended up being fractured solder joints along the rear wall on the main PCB, AND solder fractures on the round pins of the main Bridge Rectifier.



This is what I often find on the bridge.

The fractures around the output jacks aren't usually as dramatic, but close inspection with bright light & magnification will usually reveal thin hairline radial fractures around the jack's solder terminals. Seems like I spend 90 percent of my time searching for and repairing solder fractures than anything else.

Glad to hear you've got the problem on the run....just have to find all of it.

[dwmorrin]

Now that the amps appears to be functioning OK, I checked the mV across the source resistors, and looks like I have a problem here.
Amp B's P-Chan devices for instance: (all numbers in mV, red lead to one side of 0.47Ω source resistor, black to the other side of the resistor)
5
7
44
6
30

25mV +/-5mV is the magic number on the schematic.
All the resistors read about 0.50Ω using a 4 wire test in-circuit, so I'm not suspecting the resistors.
I am suspecting that the output transistors are either unmatched or were within tolerance some time ago and have drifted since.
All 4 of the MOSFET groups are showing too wide (4mV-40mV) a range of drops across the source resistors.
Probably time for a new set of MOSFETs?

[nevetslab]

I run into that often, and while the precision side of me says gotta make it all perfect, I resist the temptation. I'd check the amp at 1W and 10W output levels, and see what the voltage drop is across the Source resistors then Parts do drift over time, and the manufacturers aren't putting in matched parts....it's all bulk sourced from the semi mfgrs. We're relying on the Source resistors to help even those differences out. If the two parts that are showing as Current Hogs as above are still doing all the work, then I'd be more concerned. Do you find this on all four output stage 'halves' or just one?

You could get similar readings with brand new MosFET's too.

[Jazz P Bass]

Strange readings, indeed.

[dwmorrin]

I ended up ordering new transistors from Mouser, matched them, and I was then able to bias the amp to the spec on the schematic.

I started matching these transistors with SVT 3 Pro repairs. I am assuming that the 4 Pro requires the same. The transistors I pulled out are all color coded, so I assume they were matched to begin with. I've read multiple times on the forum that Ampeg uses matched transistors in the SVT 3. If they are indeed using unmatched transistors, that's news to me. But given the drastic improvement made in this repair, I will continue to match the devices for these amps.

I have never found a definitive "this is how you match the transistors for Ampeg MOSFET amps" post in this forum or anywhere else on the internet. I find there is quite little info on the subject in general... maybe I'm just looking in the wrong places.

What I have been using is the Nelson Pass article "How To: Matching Devices" (http://www.firstwatt.com/pdf/art_matching.pdf).
The test is for Vgs, and I've read anecdotal reports that the Ampeg spec is to match Vgs within 200mV; Pass recommends 100mV, and I've been able to get 50mV matched sets from tubes purchased from Mouser.

What I have no info on is how important it is, or how to select, the current for the Vgs test. Just plopping the IRFP240/9240 devices into Pass' schematic, I get about 250mA through each device. I don't test the devices at any other current. Does it matter? Should they be tested over a range? Since a bass guitar amp is expected to see heavy use/abuse, should the testing include a fairly high current test? The IRFP240's max specs are 20A @ room temp, 12A @ 100˚C, and the IRFP9240 max is 12A @ room temp and 7.5A @ 100˚C. Would you select a current for the Vgs test to reflect the specs of the device?

Basically, I'm just curious if the test I'm using (Nelson Pass article) is indeed appropriate for these SVT Pro amps, or if it should be modified in any way.

[MarkusBass]

Of course MOSFETs installed in SVT4Pro are matched. Most probably not by Ampeg but by the manufacturer of the transistors. They are marked with a colour dot which most probably corresponds to Vgs(off) range.
The method they use is most probably similar to the one you used. This is the reason why in the amps that just left the factory there are discrepancies in currents of the MOSFETs (because the method is not that precise). But they are in a range of 30-40 per cent and not as high as you got. It seems that 30-40 % accuracy is sufficient for this amp.
If you want to know the current for the operation, you can calculate it. You have 0.47R resistor and 25mV - it looks like 100mA test is better than 200mA.
I was also using the method you mentioned. But there is much better method if you have a working amp. You just set average current, desolder one transistor, solder three long wires, attach tested transistor to a (isolated) heatsink and measure the transistor "in-circuit". You can even do it without soldering, with clips, assuming you are very careful when doing this. Of course, tested transistor has to be connected when the amp is turned off :-). With this method you can easily get 5 per cent accuracy - much better than at the factory.
Using transistors from one batch is not enough. I usually buy 10 transistors to get 4-5 of them matched.

Mark

[dwmorrin]

Thanks for the info, Markus. I buy in the factory pack quantity from Mouser, which is 25 devices for these transistors, as it is my understanding that will give the best chances they will all be from the same factory "batch" (i.e. made on the same day, in the same factory, if lucky enough from the same hunk of silicon).
I've found that with the Nelson Pass article test, the tubes (I've maybe ordered 4 now?) have always been within 200mV of each other - but I have found that this is not good enough for the SVT amps, for I've pulled devices that wouldn't idle the same, and they would test around 200mV. The last good set was 50mV within each other. That was confusing for me because both the Pass article and the "rumored" Ampeg spec I read was 200mV. So either 200mV is no good, or my test setup (the only reasonable variable is current, I think) is no good for that number.

So you suggest matching the transistors at the idle current, which would be 0.025V/0.47Ω = 0.053A (53mA) if my ohms law math is good.

And since when I match them at 250mA they measure within the schematic spec at 53mA, I suppose that gives some evidence that the current isn't critical... although some questions still linger.

For instance, the datasheet for the IRFP240, page 3, fig 3 (http://www.irf.com/product-info/data...ta/irfp240.pdf) shows drain current plotted against Vgs for a "typical" device, and it is quite curvy and presumably a slightly different curve for each device.
This suggests to me that one could fool themselves into thinking they had matched devices by testing only at 1 current.
Maybe the ultra careful tech would measure multiple currents? (And if so, which ones? 1A? 10A?)
Or would such testing be absolutely over the top and not worth anyone's time?

[nevetslab]

SVT-4 PRO typical MosFET source current distribution

Having just gone thru this Ampeg SVT-4 PRO amp doing preventative maintenance (tightening up & finding escaped hardware, fixing solder fractures, etc), I stopped to make some bias voltage & current measurements, both @ idle, and then at 10W & 100W output levels into 8 ohms. I had already re-assembled the amp before starting in on this, so I only looked at the two upper-half channels. Gaining access to all of the two power amp circuits is a challenge with the LV P/S assembly in the way, even after removing the rear panel's upper PCB assy.

Initially, looking at Q110, Q112.....Q118 & Q210, Q212....Q218, being easy to get at, I got the variation of bias voltage across each of the 0.47 ohm/5W resistors (measuring from output buss to each MosFET Source lead). Then, I looked at the AC voltage across each of the resistors with the amp driving 10W & 100W into 8 ohm resistive load, to see what the current distribution was under load. After summing the calculated currents and computing the output power based on one-half of the output stage, the power output didn't agree with what I had set up voltage-wise (8.99V for 10W/8 ohm, and 28.3V for 100W/8 ohm). After stopping and measuring the individual 0.47 ohm Source resistors (after zeroing the Fluke 8060A DMM, none were reading 0.47 ohm, but higher, from 0.52 ohm thru 0.59 ohm. After re-calculating and using the actual values, the computed power based on each MosFET's RMS current summed together, it was considerably closer. I also connected my scope up, measuring differentially across the Source resistors (output buss vs each Source lead), the waveshape was psuedo-square-wave with variations. So, just how accurate one can get to the power output using the 10 summed currents compared to computed power using the sinewave voltage level and knowing the load & cable resistance is a bit of an instrumentation challenge. I was mostly interested in the Current Sharing under load, not the precision of calculating power output via (I^2)(R) from all the currents.

XSTR Bias V @ Idle Bias I @ Idle ACV @ 10W ACI @ 10W ACV @ 100W ACI @ 100W Source Res
Q110 29.3mV 56.3mA 62.8mV 121mA 190mV 365mA 0.52 ohm
Q112 20.5mV 37.3mA 51.2mV 93.1mA 173mV 318mA 0.55 ohm
Q114 23.6mV 47.8mA 58.0mV 105mA 185mV 336mA 0.55 ohm
Q116 25.4mV 44.6mA 59.6mV 105mA 188mV 330mA 0.57 ohm
Q118 26.9mV 49.8mA 63.0mV 117mA 195mV 361mA 0.54 ohm

Q210 25.8mV 47.8mA 55.2mV 102mA 179mV 331mA 0.54 ohm
Q212 29.8mV 52.3mA 59.6mV 105mA 188mV 330mA 0.57 ohm
Q214 36.4mV 63.9mA 65.5mV 115mA 194mV 340mA 0.57 ohm
Q216 24.3mV 41.2mA 53.0mV 89.8mA 178mV 302mA 0.59 ohm
Q218 22.8mV 40.7mA 49.3mV 88.0mA 169mV 302mA 0.56 ohm

The currents measured between the output buss and the MosFET Source terminals also take in the IR resistance of the circuit traces, which are small, but still there. Without pulling the power amp assembly again to get direct access to the 0.47 ohm resistors, this made the most sense, and did get closer to the computed power output via summed currents. I didn't make a chart of the bottom half of the power amp circuit. All the voltage measurements were made with the Fluke 8060A ACV RMS, except for the Idle Bias Voltage (DCV). Currents calculated from the measured Source Resistances.

Some images of the set-up on this Ampeg SVT-4 PRO, in constant use from our Guitar Dept Rental Inventory:




The current distribution is reasonable under drive conditions. When I had gone back to measure the bias at idle, after accumulating the data, there was a good 25-35% drop in bias voltage, a result of the heat sink temperatures.

The scope was set for DC coupling, after trimming the differential balance with the scope's 400mV square wave source. As expected, there's around 0.6V level shift before the MosFET cuts off...only conducting on one-half of the output waveform. All the waveforms across the source resistors loooked pretty much like this. Signal frequency is 400Hz. One photo shows the AC Mains Variac & Digital Power Analyzer, sitting next to an 8-ch 500W/4 ohm fan-cooled load bank

In the scope photo, the Ch 4 trace's V/Div readout should read 200mV/Div.....same as Ch 3...both channels used as the differential probe amp.