I swapped out the caps on the PC board again. This is PC14B.
The Service Manual can be found at www.hifiengine.com
All the caps were inserted with the correct polarity. The amp was working just before the swap.
I carefully swapped them ALL out with Elna Silmic II, in the correct orientation. I had some trouble with the larger caps, and could not insert them all the way in. These were C3.
After doing the work, I plugged it in and checked the DC offset.
Sadly, it was 75 volts on both channels.
No pop, click, nothing, no blown fuse! Meters do NOT deflect.
Just 75 volts DC at the outputs of both channels.
Got +/- 80 at the filter caps -- they are jensen 4-pole I put in months ago.
Some traces were lifting, as this operation had been done before at least twice, once by me.... the pcb is in bad shape. in several locations, the traces have been replaced with wire.
I'm totally lost at this point.
The output drivers appear to be ok from the service manual, the resistors R37 and R36 measure ok -- both are 146 ohms.
That leaves the output transistors. But, they too seem fine. I may be measuring wrong. Some help here would be awesome.
But what the heck? I was just changing the caps, nothing else. I did not touch the harness, all seems well.
Any clues as to the next step?
Thanks!
The Service Manual can be found at www.hifiengine.com
All the caps were inserted with the correct polarity. The amp was working just before the swap.
I carefully swapped them ALL out with Elna Silmic II, in the correct orientation. I had some trouble with the larger caps, and could not insert them all the way in. These were C3.
After doing the work, I plugged it in and checked the DC offset.
Sadly, it was 75 volts on both channels.
No pop, click, nothing, no blown fuse! Meters do NOT deflect.
Just 75 volts DC at the outputs of both channels.
Got +/- 80 at the filter caps -- they are jensen 4-pole I put in months ago.
Some traces were lifting, as this operation had been done before at least twice, once by me.... the pcb is in bad shape. in several locations, the traces have been replaced with wire.
I'm totally lost at this point.
The output drivers appear to be ok from the service manual, the resistors R37 and R36 measure ok -- both are 146 ohms.
That leaves the output transistors. But, they too seem fine. I may be measuring wrong. Some help here would be awesome.
But what the heck? I was just changing the caps, nothing else. I did not touch the harness, all seems well.
Any clues as to the next step?
Thanks!
Last edited:
The output devices are not blown. They 'being 'told' to go to +75. If any had failed you'd see fuses blowing. What is the Voltage on the collector of Q5? If it's +75 then check collector of Q4. It should be -79 - as stated on the schematic. What's the base of Q2?
You've broken the negative feedback loop and the bootstrap resistors (R17 and R18) have nothing to pull them down. Nothing really catastrophic has happened yet but you've likely left a broken trace or two. OR you may be missing the -80 supply to the board. Start measuring. In this particular case it might be safer to run on reduced Voltage with a Variac. More info will get better answers.
G²
You've broken the negative feedback loop and the bootstrap resistors (R17 and R18) have nothing to pull them down. Nothing really catastrophic has happened yet but you've likely left a broken trace or two. OR you may be missing the -80 supply to the board. Start measuring. In this particular case it might be safer to run on reduced Voltage with a Variac. More info will get better answers.
G²
Phase Linear 400
We have spare PCBs from 400s which we rebuilt using our own PCB design.
All the PLs we refurbish but keep the original PCB we do the following:
Convert to full complementary output stage
Degenerate the differential amplifier's emitters
Install a constant current source for the differential amplifier
Install base stoppers on the 40327 and 2N5415/MM4003 predrivers.
We use MJ15003/MJ15004 for PL400s and MJ1500 series for 700s
Remove the stupid input series resistor on the diff amp as it causes so much phase shift.
Change the value of the down feedback resistor to 620 ohm.
Steve Mantz
Zed Audio
We have spare PCBs from 400s which we rebuilt using our own PCB design.
All the PLs we refurbish but keep the original PCB we do the following:
Convert to full complementary output stage
Degenerate the differential amplifier's emitters
Install a constant current source for the differential amplifier
Install base stoppers on the 40327 and 2N5415/MM4003 predrivers.
We use MJ15003/MJ15004 for PL400s and MJ1500 series for 700s
Remove the stupid input series resistor on the diff amp as it causes so much phase shift.
Change the value of the down feedback resistor to 620 ohm.
Steve Mantz
Zed Audio
I'll make those tests this evening -- thanks! I feel better about it.
The PC board is getting power.
I think caps C3 may be the issue, as they have thicker leads than the others. It was a real chore to get them through the PC board, then the traces did lift.
The rest of the caps went in uneventfully, so maybe the traces broke when taking the others out?
Of course, this board has been worked on by so many other hacks, not just me, that there are lifted traces and patches all over it.
Thanks, I'll post the voltages later.
Maybe this is a good time to replace the ancient overly stiff harness?
That's a lot of work! My PC board is so beat up, it would be great to have it replaced. Any chance I can get a PL14 from you?
My amp was the latest version, with the 10K and 39K voltage divider in front of the first differential pair.
I've changed that to 15K and 59K, to be more tube friendly. When running, the DC offsets were 66 and 75 mV, the bias was .43 volts, the lowest it would go is .4 Not bad for an old amp I thought....
I don't understand your phase shift comment.... How can an input resistor cause phase shift? I and V remain in phase across a resistor.... could you please shed some light on this? thanks.
Check all of the wires to and from the PCB, since you've been moving it around a wire may have broken either at the board or at the output section. Next check the wires to the bias transistors, red, brown and black. The last 400 I rebuilt was working fine until I screwed all back together and it didn't work. Disassembled one more time and found the leads to the bias transistors had broken.
Craig
Craig
Thanks, I will check each one, especially the bias. Could I measure the bias voltage to make sure? If the bias checks out, then the feedback loops on the PC board are broken.
I think I should replace all of the wires while I am at this... I'll dip the ends
into flux, tin them, dip into flux again and use the existing solder on the pads to connect. Those pads are really close together.
I think I should replace all of the wires while I am at this... I'll dip the ends
into flux, tin them, dip into flux again and use the existing solder on the pads to connect. Those pads are really close together.
I put through holes on the PCB, and rewired the harness with new wires. There was no change. Only the B+ and B- wires and the bias transistor Q6 are still using original wire. I am not surprised, but at least there is new wire
The base of the bias transistors Q6 shows 75 volts on either channel. It is supposed to show 0 according to the test voltages.
I will check continuity to the caps I changed once more, then provide the test voltages. But, I can't do anymore on this until Sunday night.
At that point, I'll post some hard data.
Is it safe to run the amp like this? I mean, no load, no input but 75 volts at the output? I don't want to blow stuff up....
The base of the bias transistors Q6 shows 75 volts on either channel. It is supposed to show 0 according to the test voltages.
I will check continuity to the caps I changed once more, then provide the test voltages. But, I can't do anymore on this until Sunday night.
At that point, I'll post some hard data.
Is it safe to run the amp like this? I mean, no load, no input but 75 volts at the output? I don't want to blow stuff up....
PL400
Email me at zedaudio@aol.com to discuss the PCB.
Follow the schematic as to what the input network is.
Early 400s had a series 1K with 220pF and a gain of 23 times, this is OK.
The next generation had NPN diff amps with an intermediate PNP diff amp, used 2K7 with 220pF but had attenuation of a 2K7 with a 10K (0.787x) and then feedback parts were 10K with 270R down = gain of 38x. Final gain was therefore 29.9x - in my opinion too high. Why atteniate on the inut and then gain up later makes no sense.
The final 3 generations of the quasi comp versions used a 10K series input with a 39K bias and a 220pF to ground. So the phase shift is caused by the 10K/220pF and the attenuation by the 10K/39K (0.795).
At 16KHz this 10K/220pF is down 0.22dB, at 20KHz down 0.3dB and down 3dB at 74KHz.
They claim response out to 250KHz, I think not.
Phase is lagging 16 degrees at 20KHz, they claim 12 degress for the total amplifier and this 16 degrees does not take into account the propogation delay within the amplifier.
At 20KHz the phas elag of the whole amplifier is 35 degrees!
To make these amplifiers sound halfway decent requires many modifications. We have done many PL400 and 700s with good success.
The most expensive parts are the new main capacitors (22,000mfd 100v) and the power devices (Either MJ15003/4 for PL400 and MJ15024/MJ15025 or MJ21193/4 for 700s.
The grounding of all PLinear amps regardless of their vintage is really bad and by simple rewriing the THD of these amps drops by more than a factor of 5x. Just look at the peice of wire from the RCA ground to the right channel speaker ground...a really bad move.
The power devices share power rail wiring and so massive intermodulation distortion occurs.
A typical rebuild of one of these 400s takes me about 5 hours which includes stripping the amplifier completely, cleaning metal parts, replacing the RCA sockets with Tiffany parts, ne wbinding posts (Requires drilling out the existing holes), bolting the new power devices, rewiring the complete amplifier and final testing.
I also designed and had made a new PCB for the 400 and 700s. This is a fully complementary drive board which is vastly superior to the originals.
I also add surge protection for the large 22,000mfd caps and a relay for DC and turn on/off mute.
The last iteration of PL400/700 even though these had complementary output stages was just as bad. That LF356/LF351 opamp upfront was really bad.
The fact that no diff amp emitter degeneration was ever used in any PL amps made them really unstable, no base stoppers in the drivers and of course no true constant current sources for the diff amp and VAS stages.
The VI limiters with those germanium transistors were "tweeter eaters".
Steve Mantz
Zed Audio Corp.
Email me at zedaudio@aol.com to discuss the PCB.
Follow the schematic as to what the input network is.
Early 400s had a series 1K with 220pF and a gain of 23 times, this is OK.
The next generation had NPN diff amps with an intermediate PNP diff amp, used 2K7 with 220pF but had attenuation of a 2K7 with a 10K (0.787x) and then feedback parts were 10K with 270R down = gain of 38x. Final gain was therefore 29.9x - in my opinion too high. Why atteniate on the inut and then gain up later makes no sense.
The final 3 generations of the quasi comp versions used a 10K series input with a 39K bias and a 220pF to ground. So the phase shift is caused by the 10K/220pF and the attenuation by the 10K/39K (0.795).
At 16KHz this 10K/220pF is down 0.22dB, at 20KHz down 0.3dB and down 3dB at 74KHz.
They claim response out to 250KHz, I think not.
Phase is lagging 16 degrees at 20KHz, they claim 12 degress for the total amplifier and this 16 degrees does not take into account the propogation delay within the amplifier.
At 20KHz the phas elag of the whole amplifier is 35 degrees!
To make these amplifiers sound halfway decent requires many modifications. We have done many PL400 and 700s with good success.
The most expensive parts are the new main capacitors (22,000mfd 100v) and the power devices (Either MJ15003/4 for PL400 and MJ15024/MJ15025 or MJ21193/4 for 700s.
The grounding of all PLinear amps regardless of their vintage is really bad and by simple rewriing the THD of these amps drops by more than a factor of 5x. Just look at the peice of wire from the RCA ground to the right channel speaker ground...a really bad move.
The power devices share power rail wiring and so massive intermodulation distortion occurs.
A typical rebuild of one of these 400s takes me about 5 hours which includes stripping the amplifier completely, cleaning metal parts, replacing the RCA sockets with Tiffany parts, ne wbinding posts (Requires drilling out the existing holes), bolting the new power devices, rewiring the complete amplifier and final testing.
I also designed and had made a new PCB for the 400 and 700s. This is a fully complementary drive board which is vastly superior to the originals.
I also add surge protection for the large 22,000mfd caps and a relay for DC and turn on/off mute.
The last iteration of PL400/700 even though these had complementary output stages was just as bad. That LF356/LF351 opamp upfront was really bad.
The fact that no diff amp emitter degeneration was ever used in any PL amps made them really unstable, no base stoppers in the drivers and of course no true constant current sources for the diff amp and VAS stages.
The VI limiters with those germanium transistors were "tweeter eaters".
Steve Mantz
Zed Audio Corp.
I also changed the grounding, in preparation for changing the input jacks. But, all I did (I think) was connect the signal ground on each channel to the output ground to free up the RCA board.
The original board has the wires attached at a point between the RCA's, then a wire was connected to the Right channel output ground. All I did, was move those wires closer to that jack, so that should not have changed anything.
I assume that connecting the signal ground to hot would actually fry the output transistors instantly and/or blow fuses.
Any idea what failure modes I could induce to mess that stuff up?
I will be checking voltages later, so long as I can find the time.... my wife broke her leg on the weekend, and I'm sure she'd prefer my company
The original board has the wires attached at a point between the RCA's, then a wire was connected to the Right channel output ground. All I did, was move those wires closer to that jack, so that should not have changed anything.
I assume that connecting the signal ground to hot would actually fry the output transistors instantly and/or blow fuses.
Any idea what failure modes I could induce to mess that stuff up?
I will be checking voltages later, so long as I can find the time.... my wife broke her leg on the weekend, and I'm sure she'd prefer my company
Any reason that the high gain TIS97's can't be subbed with a 2N2222A or 2N3904?
They are lower gain devices. Won't that alleviate some of the issues with the TIS97 differential pair, and allow for the input resistor to have it's value significantly decreased -- thus raising the bandwidth? Or is the bandwidth intentionally lowered due to instability?
They are lower gain devices. Won't that alleviate some of the issues with the TIS97 differential pair, and allow for the input resistor to have it's value significantly decreased -- thus raising the bandwidth? Or is the bandwidth intentionally lowered due to instability?
Why the meters don't show full power out is beyond me.... One meter was showing full power, but it stopped doing that -- I assume that there was a short in the meter wiring harness, that cleared when the harness was unstressed.
I reflowed the solder on each pad. I caused a problem but fixed that -- the fix of a broken trace went bad, but I repaired it.
Each pad is connected to it's neighboring pad. There does not appear to be any broken traces on the board.
So, it can only be a part.
I got one test voltage showing that the emitters of differential pair Q1 and Q2 show 3.2/9.8 for L/R channel respectively, as opposed to -0.6.
I can get more voltages, but I am not well versed enough to identify why they are the way they are.... Any help would be most appreciated.
Thanks in advance.
I reflowed the solder on each pad. I caused a problem but fixed that -- the fix of a broken trace went bad, but I repaired it.
Each pad is connected to it's neighboring pad. There does not appear to be any broken traces on the board.
So, it can only be a part.
I got one test voltage showing that the emitters of differential pair Q1 and Q2 show 3.2/9.8 for L/R channel respectively, as opposed to -0.6.
I can get more voltages, but I am not well versed enough to identify why they are the way they are.... Any help would be most appreciated.
Thanks in advance.
Last edited:
Having time away from the amp, I decided to learn something about BJT differential pairs.
Given that the voltage at the emitters of Q1 and Q2 is 6 volts, it appears that the V- is disconnected from them. That is the first thing to check upon my return to the city...
If there is no continuity from the B- to the emitters of Q1 and Q2, then I've found what could be the only problem....
Given that the voltage at the emitters of Q1 and Q2 is 6 volts, it appears that the V- is disconnected from them. That is the first thing to check upon my return to the city...
If there is no continuity from the B- to the emitters of Q1 and Q2, then I've found what could be the only problem....
Found some diodes that need replacement, but not all in the same channel.
D3, D4 and D9. These will be replaced tomorrow.
Q3 and Q4 were both MPA093, and were changed to a matched pair.
Q5 was replaced.
Q1/Q2 were removed from the left channel and hfe was tested -- one was at 139 the other 567. They were replaced with matched 2n2222 of gain 164.
The Q1/Q2 differential pairs have 9.8 volts on both emitters and collectors - BOTH CHANNELS.
Continuity was tested on the feedback loop on the left channel. It is fine from the output to the base of Q2 -- no breaks in the global feedback loop. Right channel was not tested.
There is continuity from the transistors to neighboring pads on all channels.
There is continuity from the PCB to the bias transistors off board.
The bias Potentiometers read 4.6K each. I have not checked the wipers.
There is no change in status, even though the transistors that are listed as causing this problem have been changed. Still 75 volts at the outputs.
I have checked the capacitor changes over and over again. They appear OK.
Does the service manual show proper polarity for all caps?
I am at a complete loss, except for possible error in polarity in the documentation. It is possible that the manual is wrong. The larger caps I have removed are showing a sticky substance on them, as if they are leaky. These caps were only used a couple weeks.
All resistors measure fine in circuit.
Any help would be most appreciated
D3, D4 and D9. These will be replaced tomorrow.
Q3 and Q4 were both MPA093, and were changed to a matched pair.
Q5 was replaced.
Q1/Q2 were removed from the left channel and hfe was tested -- one was at 139 the other 567. They were replaced with matched 2n2222 of gain 164.
The Q1/Q2 differential pairs have 9.8 volts on both emitters and collectors - BOTH CHANNELS.
Continuity was tested on the feedback loop on the left channel. It is fine from the output to the base of Q2 -- no breaks in the global feedback loop. Right channel was not tested.
There is continuity from the transistors to neighboring pads on all channels.
There is continuity from the PCB to the bias transistors off board.
The bias Potentiometers read 4.6K each. I have not checked the wipers.
There is no change in status, even though the transistors that are listed as causing this problem have been changed. Still 75 volts at the outputs.
I have checked the capacitor changes over and over again. They appear OK.
Does the service manual show proper polarity for all caps?
I am at a complete loss, except for possible error in polarity in the documentation. It is possible that the manual is wrong. The larger caps I have removed are showing a sticky substance on them, as if they are leaky. These caps were only used a couple weeks.
All resistors measure fine in circuit.
Any help would be most appreciated
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.