Thanks Sangram, your suspicion was right. I accidentally measured the sources due to misunderstanding the pinouts. I got supply voltage at the JFET drains and on both sides of each pot. For some unknown reason, the JFET's don't seem to be heating up now either, now that I have unmounted and remounted the PCB's and a few components. I decided to go through the bias setting procedure again, and now it seems fine! Some of my keratherm's got really mangled and I replaced them. If any of my diodes were contacting the heatsink, could it have caused the symptoms I described?
I'm currently listening to sweet, sweet, music. I'm getting some hum, so I will try Nelson's grounding rectifier/thermistor circuit. I will also build the store protection boards to minimize the chance of blowing up my now hard-won effort.
Thanks!
-Chris
Folks:
I think I'm at the verge of making a dumb mistake and hope you can provide some guidance.
The front-end board on one of my F5 Turbo V3 amps has been replaced with all-new components and 470pF caps have been added across R7 and R10. The N-channel MOSFETs and the 47R resistors on the output boards have also been replaced (the latter were upgrades from 0.25W to 1.0W units). The amp should be good to go, but I'm still suspicious for two reasons:
I'm no diagnostician. Can anyone suggest my next steps?
Many thanks,
Scott
I think I'm at the verge of making a dumb mistake and hope you can provide some guidance.
The front-end board on one of my F5 Turbo V3 amps has been replaced with all-new components and 470pF caps have been added across R7 and R10. The N-channel MOSFETs and the 47R resistors on the output boards have also been replaced (the latter were upgrades from 0.25W to 1.0W units). The amp should be good to go, but I'm still suspicious for two reasons:
1. With the amp and dim bulb tester on, voltage at both rails is about 24VDC. When powering down the amp (while connected to the dim bulb tester), voltage on the positive rail drops quickly. Voltage on the negative rail, on the other hand, drops very slowly. By the time the positive rail gets to 2 VDC, the negative rail is about 18 VDC.
2. With the amp and dim bulb tester on and an 8 ohm load on the amp's output, the relays on the speaker protection board cycle on and off.
2. With the amp and dim bulb tester on and an 8 ohm load on the amp's output, the relays on the speaker protection board cycle on and off.
I'm no diagnostician. Can anyone suggest my next steps?
Many thanks,
Scott
Hi Scott,
Since you have another working mono check for similar behavior on that one.
From what I understand you haven't really found what caused your problem.
Were the cascode transistors fried? What bias were you running the mono at?
Are the values of all other parts as per the F5Turbo V3 schematic,? What R7 thru R10 values do you have?
nash
Since you have another working mono check for similar behavior on that one.
From what I understand you haven't really found what caused your problem.
Were the cascode transistors fried? What bias were you running the mono at?
Are the values of all other parts as per the F5Turbo V3 schematic,? What R7 thru R10 values do you have?
nash
I assume you did not have an smoke or overheating parts on the dim bulb tester? If so, that’s a good sign. The power supply decay is telling you that the negative side is not drawing much current compared to the + side. Do you have the bias pots turned down all the way as you normally would when starting the biasing procedure? You need to set up as if you were bringing the amp up for the first time. Go back to the manual or the build guide and follow the steps for adjustment. With the bias pots set at 0 ohms you should not have any voltage across any of the output stage source resistors.
If you are running cascodes - check for the proper voltages at the emitters of the cascode transistors.
These are just a couple of basic checks to try.
If you are running cascodes - check for the proper voltages at the emitters of the cascode transistors.
These are just a couple of basic checks to try.
nashbap:
Yes, the other monoblock behaves as it should; within 8-10 seconds, both rails drop to under 2 VDC after it is shut down. And you're right -- I did not figure out what caused the amp's failure; I lack the expertise to diagnose causation, so I've focused on examining the parts (both active and passive) and replacing anything that has blown or shows signs of wear.
I haven't tested the cascode transistors. The amp's rails are 44.2 VDC +/- and bias was originally set at +0.311 and -0.360 VDC (0.02 VDC offset). However, that was 6.5 years ago and I confess the bias hasn't been checked in several years. Is it possible that there was enough drift to cause a failure? My speakers are fine (Elsinores now, but the Andras that had been in that system are still working fine elsewhere in the house).
The values are all consistent with the F5T V3 specifications. R7-R10 are all 220R 3W Vishay RS series resistors.
Bfpca:
Also correct -- no smoke with the dim bulb tester. I will double-check but am confident that the bias pots were both dialed all the way down.
I have the other monoblock on my bench at the moment; I upgraded about a dozen resistors and added bypass caps to R7 and R10 and am now setting its bias. Early in the process, it is behaving as expected. I'll report back on the cascode voltages when I complete biasing the working amp.
Thanks for the counsel so far.
Regards,
Scott
Yes, the other monoblock behaves as it should; within 8-10 seconds, both rails drop to under 2 VDC after it is shut down. And you're right -- I did not figure out what caused the amp's failure; I lack the expertise to diagnose causation, so I've focused on examining the parts (both active and passive) and replacing anything that has blown or shows signs of wear.
I haven't tested the cascode transistors. The amp's rails are 44.2 VDC +/- and bias was originally set at +0.311 and -0.360 VDC (0.02 VDC offset). However, that was 6.5 years ago and I confess the bias hasn't been checked in several years. Is it possible that there was enough drift to cause a failure? My speakers are fine (Elsinores now, but the Andras that had been in that system are still working fine elsewhere in the house).
The values are all consistent with the F5T V3 specifications. R7-R10 are all 220R 3W Vishay RS series resistors.
Bfpca:
Also correct -- no smoke with the dim bulb tester. I will double-check but am confident that the bias pots were both dialed all the way down.
I have the other monoblock on my bench at the moment; I upgraded about a dozen resistors and added bypass caps to R7 and R10 and am now setting its bias. Early in the process, it is behaving as expected. I'll report back on the cascode voltages when I complete biasing the working amp.
Thanks for the counsel so far.
Regards,
Scott
Hi Scott,
I do feel it is important to first check whether the cascode transistors blew and that caused your problem or was it something else. What transistors are you using?
Additionally, check the grounds thoroughly. In my 45v monoblock build, I have separate wires from PSU star ground to each output bank, and a separate wire to the FE board. How are yours routed? A poor ground will cause an effect similar to what you experienced.
Best,
nash
I do feel it is important to first check whether the cascode transistors blew and that caused your problem or was it something else. What transistors are you using?
Additionally, check the grounds thoroughly. In my 45v monoblock build, I have separate wires from PSU star ground to each output bank, and a separate wire to the FE board. How are yours routed? A poor ground will cause an effect similar to what you experienced.
Best,
nash
Good to hear. It's not fun when an amp gives trouble, specially one as nice as this.
Scott, you need to run voltage checks on the front end board, and I wouldn't connect the power devices board to it until you are confident of its behaviour. start by checking the JFET current across the source resistors, should be about 60-80mV for a JFET current of 6.5-8mA. Next, with the trimpots measured zero, you need to have zero volts between the MOSFET's S and G pins (That's the collector of the cascode to the rail on each polarity).
Keeping your meter connected to these points, start moving the trimmer and you should get a reading up to 5-6V at maximum trimpot setting. Back it off again till you measure zero. This means the board is biasing up properly and you're ready to connect the output boards/add the MOSFETs.
I typically do this almost every build out of insane compulsion, just to ensure I've got it right.
Folks:
Close but not quite there yet. I've built an entirely new front end board for my blown F5T V3 mono block and replaced all of its MOSFETs. In addition, I've added 470 pF caps at R7 and R10 to control oscillation, changed out R3 and R4 for 10R / 3W resistors and increased the size of the gate resistors on the output boards (R134-R16) to 1W. The rails are about +/- 44.7 VDC unloaded.
The problem is in the biasing of the amp. P1 is a 25 turn 5k pot and, before it was installed, it tested properly (i.e., ranging from close to 0 to about 5k). But adjusting the bias on the positive side of the amp isn't proceeding well -- with P1 turned all of the way up, I'm measuring just 0.161 mV on the P-side output boards' test points. P2 isn't turned nearly as far (maybe half-way?). Before I pull and test the positive cascode (2SC4793) or the jfets, can anyone tell me if I should be looking at something else? I'd like to avoid having to yank parts out unnecessarily.
Regards,
Scott
Close but not quite there yet. I've built an entirely new front end board for my blown F5T V3 mono block and replaced all of its MOSFETs. In addition, I've added 470 pF caps at R7 and R10 to control oscillation, changed out R3 and R4 for 10R / 3W resistors and increased the size of the gate resistors on the output boards (R134-R16) to 1W. The rails are about +/- 44.7 VDC unloaded.
The problem is in the biasing of the amp. P1 is a 25 turn 5k pot and, before it was installed, it tested properly (i.e., ranging from close to 0 to about 5k). But adjusting the bias on the positive side of the amp isn't proceeding well -- with P1 turned all of the way up, I'm measuring just 0.161 mV on the P-side output boards' test points. P2 isn't turned nearly as far (maybe half-way?). Before I pull and test the positive cascode (2SC4793) or the jfets, can anyone tell me if I should be looking at something else? I'd like to avoid having to yank parts out unnecessarily.
Regards,
Scott
Hi, all!
Just to hear you guys out wrt dual mono V2 configurations: if i was to use two 600va transformers side by side, should I worry about distance to FE’s in the 5U chassis? Anyone tried and succeded/not succeded?
I’m struggling to decide whether to go for dual mono or not. Have both options, but PSU boards are allready finished, and must be recapped if I use only one PSU. Just to be clear: 1 transformer in each corner close to front plate, FE’s in the middle of heatsinks.
Regards,
Andreas
Just to hear you guys out wrt dual mono V2 configurations: if i was to use two 600va transformers side by side, should I worry about distance to FE’s in the 5U chassis? Anyone tried and succeded/not succeded?
I’m struggling to decide whether to go for dual mono or not. Have both options, but PSU boards are allready finished, and must be recapped if I use only one PSU. Just to be clear: 1 transformer in each corner close to front plate, FE’s in the middle of heatsinks.
Regards,
Andreas
Last edited:
Scott,
You first need to turn the pot more than halfway. Using the turns position to intuit front end operation is not good practice. Assuming you do, and don't get any results, you can follow some additional steps below.
The first thing to do is check JFET source current across the 10ohm resistors. As foranobject discovered, you should have between 60 and 90mV across them, indicating the front end is working properly - corresponding to JFET current between 6 and 9mA. At 6mA, a 1k lumped drain resistance will allow 6V of bias to the output device, which should be enough for a few amperes.
The next step is to check the drain voltages of the JFETs to ensure the cascodes are properly dropping voltages. sometimes a mix-up between resistor values in the cascode base network will cause zero drain voltages (or too little). You should get identical values between the two JFETs for this test, or close to it.
As I've mentioned, you need not connect output boards while running the voltage/current checks on the front end. If you have done this, you can test the front end operation by check the bias voltage between pots set to zero, maximum and halfway. About 6V of bias should be more than enough. This measurement is taken between cascode collector and the supply rail. This test cannot be conducted with the output boards connected.
If you find you're not getting enough bias voltage, 5V or less, you should go ahead and change R5 and R6. With a 5k pot, 1K drain resistor and 6mA of JFET current, you will end up with just about 5V, which may not be enough.
You first need to turn the pot more than halfway. Using the turns position to intuit front end operation is not good practice. Assuming you do, and don't get any results, you can follow some additional steps below.
The first thing to do is check JFET source current across the 10ohm resistors. As foranobject discovered, you should have between 60 and 90mV across them, indicating the front end is working properly - corresponding to JFET current between 6 and 9mA. At 6mA, a 1k lumped drain resistance will allow 6V of bias to the output device, which should be enough for a few amperes.
The next step is to check the drain voltages of the JFETs to ensure the cascodes are properly dropping voltages. sometimes a mix-up between resistor values in the cascode base network will cause zero drain voltages (or too little). You should get identical values between the two JFETs for this test, or close to it.
As I've mentioned, you need not connect output boards while running the voltage/current checks on the front end. If you have done this, you can test the front end operation by check the bias voltage between pots set to zero, maximum and halfway. About 6V of bias should be more than enough. This measurement is taken between cascode collector and the supply rail. This test cannot be conducted with the output boards connected.
If you find you're not getting enough bias voltage, 5V or less, you should go ahead and change R5 and R6. With a 5k pot, 1K drain resistor and 6mA of JFET current, you will end up with just about 5V, which may not be enough.
Sangram:
My last post failed to mention that I had followed your earlier guidance -- I should have been clearer. Mea culpa.
AudioSan:
Understood.
All:
Even with the excellent support, I managed to screw things up. The rebuilt amp tested okay, bias was adjusted and I let it run in for a few hours on my bench. But within 5 minutes after installing it in my basement system there was a pop, some magic smoke and a crushing sense of loss. I have been too dispirited to dive into the amp over the last week.
Given that I had replaced all of the active components (other than the diodes on the four output boards) and many of the passive components, I'm at something of a loss. It's probably best if I start again from scratch: although I'm not looking forward to it, I'm anticipating a complete rebuild. I've ordered AI203 ceramic insulator pads to replace all of the Keratherm Red pads I had been using (a bit of a waste, since all of the recently replaced MOSFETs also got new Keratherms) for both performance and longevity reasons (and on the off-chance that one of the Keratherm pads under the diodes, which were not checked, has fallen apart).
So I'll be going dark for a bit, until critical parts start arriving. Thanks for the support so far -- I'll undoubtedly need some hand-holding once the rebuild begins. And extra thanks to vdi_nenna, who kindly loaned me one of his amps on a "temporary" basis (good luck getting it back).
Regards,
Scott
My last post failed to mention that I had followed your earlier guidance -- I should have been clearer. Mea culpa.
AudioSan:
Understood.
All:
Even with the excellent support, I managed to screw things up. The rebuilt amp tested okay, bias was adjusted and I let it run in for a few hours on my bench. But within 5 minutes after installing it in my basement system there was a pop, some magic smoke and a crushing sense of loss. I have been too dispirited to dive into the amp over the last week.
Given that I had replaced all of the active components (other than the diodes on the four output boards) and many of the passive components, I'm at something of a loss. It's probably best if I start again from scratch: although I'm not looking forward to it, I'm anticipating a complete rebuild. I've ordered AI203 ceramic insulator pads to replace all of the Keratherm Red pads I had been using (a bit of a waste, since all of the recently replaced MOSFETs also got new Keratherms) for both performance and longevity reasons (and on the off-chance that one of the Keratherm pads under the diodes, which were not checked, has fallen apart).
So I'll be going dark for a bit, until critical parts start arriving. Thanks for the support so far -- I'll undoubtedly need some hand-holding once the rebuild begins. And extra thanks to vdi_nenna, who kindly loaned me one of his amps on a "temporary" basis (good luck getting it back).
Regards,
Scott
Best wishes and good vibes heading your way to get your amp up and running.
Scott, that's really sad.
My usual recommendation for continuously failing F5T amps is to either remove the diodes, or reduce the source resistors of the MOSFET and adjust the bias according to the new values. I know the amp is no longer a F5T, but I know it works and works well.
Good luck whenever you get to fixing it.
My usual recommendation for continuously failing F5T amps is to either remove the diodes, or reduce the source resistors of the MOSFET and adjust the bias according to the new values. I know the amp is no longer a F5T, but I know it works and works well.
Good luck whenever you get to fixing it.
Hi All,
New member here. I've been researching to make an F5 Turbo V3 for a while now, and have a question regarding the active devices. From what I can tell, to make two V3 monoblocks, I would need two orders of the "F-5T V1/2 TRANSISTOR KIT" from the store? Is there a reason why it does not specify usage for the turbo V3?
Other than that, I am planning on ordering two sets of turbo pcbs and 4 power supply boards (to achieve a cap bank as specified in the original pdf). Anyways, thanks for the help
New member here. I've been researching to make an F5 Turbo V3 for a while now, and have a question regarding the active devices. From what I can tell, to make two V3 monoblocks, I would need two orders of the "F-5T V1/2 TRANSISTOR KIT" from the store? Is there a reason why it does not specify usage for the turbo V3?
Other than that, I am planning on ordering two sets of turbo pcbs and 4 power supply boards (to achieve a cap bank as specified in the original pdf). Anyways, thanks for the help
Sorry to hear your experience.
It has helped, by testing if the output transistors have shorted to case or not. At least once, the short was intermittent -- little more pressure on the output device and it would show as a short. This might have happened to you. Keratherm is fragile.
I agree with the use of the ceramic insulators for that reason.
It has helped, by testing if the output transistors have shorted to case or not. At least once, the short was intermittent -- little more pressure on the output device and it would show as a short. This might have happened to you. Keratherm is fragile.
I agree with the use of the ceramic insulators for that reason.
Hi All,
I need a sanity check on bias voltages. I have built a pair of monoblock F5 Turbo V3s using 30V secondaries and am getting about 44V DC unloaded and 39V loaded. In setting the bias I have placed DC volt meters on all four output boards and am getting more variability than I feel comfortable with. On the "P" side I read 350mv and 288mv. The "N" side is 273mv and 297mv. They will vary somewhat over time but the ratio remains about the same. Offset stays within +5mv. Seems like I have too much variation between boards with 60+ mv on the P side and 30mv on the right. What should be the maximum variation that would be acceptable?
Thanks,
Roy
I need a sanity check on bias voltages. I have built a pair of monoblock F5 Turbo V3s using 30V secondaries and am getting about 44V DC unloaded and 39V loaded. In setting the bias I have placed DC volt meters on all four output boards and am getting more variability than I feel comfortable with. On the "P" side I read 350mv and 288mv. The "N" side is 273mv and 297mv. They will vary somewhat over time but the ratio remains about the same. Offset stays within +5mv. Seems like I have too much variation between boards with 60+ mv on the P side and 30mv on the right. What should be the maximum variation that would be acceptable?
Thanks,
Roy