I thought V- LED is the LED on the right side the board.
I measured V- to Ground at the amp board test points in both channels, next to the soldering points for the same.
Yes there was a lot of heat generated when running before connecting source and speakers. Atleast like a 100W incandescent light bulb.
I measured V- to Ground at the amp board test points in both channels, next to the soldering points for the same.
Yes there was a lot of heat generated when running before connecting source and speakers. Atleast like a 100W incandescent light bulb.
You're right, the LED on the right does seem to be connected to the negative supply.
I would think there's something wrong with your PSU. If the amp circuit had a defect and would load the PSU to the point where the voltage drops to -3V, surely the fuse would pop as this would require a tremendous amount of current.
Heat is kinda good, that means the active devices are switched on and are conducting current. Iirc, VFETs mostly fail open, which would mean no current and therefore no heat. So I'd assume the amp worked at one point.
The single most important voltages of the amp are T16 and T17, the bias voltages for the VFETs. If they are not present (or too close to zero) the VFETs will run wide open (normally on devices) and burn up. I would therefore advise against switching on the amp with the VFETs installed until the PSU is checked thoroughly. No negative supply voltage means no negative bias which means Q9 is threatened.
If it were my amp, I would disconnect the PSU from both amp boards and check it independently. Do you know how to do that safely?
And, as always, please post pics.
I would think there's something wrong with your PSU. If the amp circuit had a defect and would load the PSU to the point where the voltage drops to -3V, surely the fuse would pop as this would require a tremendous amount of current.
Heat is kinda good, that means the active devices are switched on and are conducting current. Iirc, VFETs mostly fail open, which would mean no current and therefore no heat. So I'd assume the amp worked at one point.
The single most important voltages of the amp are T16 and T17, the bias voltages for the VFETs. If they are not present (or too close to zero) the VFETs will run wide open (normally on devices) and burn up. I would therefore advise against switching on the amp with the VFETs installed until the PSU is checked thoroughly. No negative supply voltage means no negative bias which means Q9 is threatened.
If it were my amp, I would disconnect the PSU from both amp boards and check it independently. Do you know how to do that safely?
And, as always, please post pics.
I sincerely appreciate your responses Rodeodave. Thanks.
I took apart the power supply from the amp boards and it appears that there is no V- . It must have been ok when I did the initial power up and biasing.
I'll debug and hope to get this going. When I am at that point, how would I test this under load, without connecting to the amp boards ? I could use some pointers.
I took apart the power supply from the amp boards and it appears that there is no V- . It must have been ok when I did the initial power up and biasing.
I'll debug and hope to get this going. When I am at that point, how would I test this under load, without connecting to the amp boards ? I could use some pointers.
I would start with inspecting solder joints. Reflow the suspicions-looking ones (or all of them) with a hot itron and new solder. PSU boards, especially ones with ground planes, can be tricky to solder.
One channel of an operational VFET amp should draw 1A of current (dropping 100mV over R32) per polarity, 2A per polarity for stereo. Loading one polarity of the PSU for such a current draw would require a 28V/2A=14R resistor that can dissipate something like 60W. Not very handy. However, to get an idea how the PSU behaves under load, you could just hook up a lightbulb (incandescent type). A 100W 110V lightbuld will have something like 120R, a 300W might be as low as 40R (that's for a hot filament at 110V), while the cold resistance should be around 10R for the 100W (see Say Goodbye To The Incandescent Lightbulb (as We Know It) | Fluke). Easy to get a hold of, doesn't require additional heatsinking.
One channel of an operational VFET amp should draw 1A of current (dropping 100mV over R32) per polarity, 2A per polarity for stereo. Loading one polarity of the PSU for such a current draw would require a 28V/2A=14R resistor that can dissipate something like 60W. Not very handy. However, to get an idea how the PSU behaves under load, you could just hook up a lightbulb (incandescent type). A 100W 110V lightbuld will have something like 120R, a 300W might be as low as 40R (that's for a hot filament at 110V), while the cold resistance should be around 10R for the 100W (see Say Goodbye To The Incandescent Lightbulb (as We Know It) | Fluke). Easy to get a hold of, doesn't require additional heatsinking.
All 4 filter resistors in the power supply for V- had blown. Not sure why, yet. They were 3W, 1.1 Ohm. I am using a diyaudio v3 psu. 2 fuses, 1.5A each. 2 fuses of 500mA each did not survive initial turn on. The transformer has a CL60 on each input pair.
I will try an incandescent light bulb when I am ready to test the power supply.
Thanks for the interesting reference.
I will try an incandescent light bulb when I am ready to test the power supply.
Thanks for the interesting reference.
Did you mean 0.1 ohm?
If not the 1.1 ohm would disapate close to its rating in constant current flow and exceed its wattage rating on turn-on.
Where are you using 500mA fuses? If for the mains fuse, it is way too small, and way too small for anywhere else.
The transformer is what the main fuse is protecting. Use what Nelson suggests in his article.
Rush
Hi Stanney,
The vfets I bought from acroman did not contain such markings.
It's possible they represent Vgs values, but at what current?
I guess you're trying to figure out where to get the initial Vgs voltage
for the vfets? The exact value is not important, only that it is high enough
so the vfets conduct minimal current.
Are you able to do some measurements on your vfets?
The vfets I bought from acroman did not contain such markings.
It's possible they represent Vgs values, but at what current?
I guess you're trying to figure out where to get the initial Vgs voltage
for the vfets? The exact value is not important, only that it is high enough
so the vfets conduct minimal current.
Are you able to do some measurements on your vfets?
Hello Dennis,
Thanks for getting back.
I bought them from two sources. Acroman and through the DIYAUDIO swap meet.
Hasnt helped moving house twice so they are a little mixed up.
I am trying to set the initial voltage. I guess I could set at 10v if it’s not important.
I did try with the method described in the forums. They generated so much heat I backed off. Was hoping I might find someone experienced with a curve tracer.
Thanks for getting back.
I bought them from two sources. Acroman and through the DIYAUDIO swap meet.
Hasnt helped moving house twice so they are a little mixed up.
I am trying to set the initial voltage. I guess I could set at 10v if it’s not important.
I did try with the method described in the forums. They generated so much heat I backed off. Was hoping I might find someone experienced with a curve tracer.
Ouch. The BOM said .47 - 1 ohm, 3W. Can you please tell me how to calculate the size of these resistors? I could populate all 7 with 1.1 .
At first I used 500mA fuses at mains switch, before the transformer. Newbie mistake. 500mA each on L and N does not equal 1A protection. Those blew. Is my understanding of this correct?
Sorry I wasn't being clear. The exact initial value is not important as long as it is high enough to make sure the vfets are not passing much current.
In your case, someone might have done some measurements. If they represent
Vgs for something like 0.5A or 1.0A then setting the initial voltage at 10V will give
way too high a current. So please don't do that until we get some more information.
Do you have the link to the setup you used when you tried measuring?
In your case, someone might have done some measurements. If they represent
Vgs for something like 0.5A or 1.0A then setting the initial voltage at 10V will give
way too high a current. So please don't do that until we get some more information.
Do you have the link to the setup you used when you tried measuring?
Last edited:
Well, yes populate all parrallel 1.1 resistors in the power supply board to share the current and reduce the resistance. To calculate the wattage dissipated use Ohms Law. You probably need to do some research so you understand the basic formulas for Ohms Law: V=I*R ; I=V/R; R=V/I
V = Voltage in volts
I = Current in amps
R = Resistance in ohms
Wattage = V*A
When you parallel equal resistors the resistance is hafted. But the formula is
1/R + 1/R + 1/R (etc.) = 1/Rt (Rt is total R) The formula is great for finding paralleled resistors of unequal valves or an odd number of resistors of the same value.
1.1R times 7 is .157 ohms. ((1/1.1)*7) = 1/Rt
I don't know the VA of your transformer, so review the original power supply schematic and start with the recommenced fuse value.
Rush
I don't know how to calculate the peak inrush current, especially with a CL60 inline.
If you go to the 1st post in this tread, 6L6 has posted the power supply schematic from Nelson's article. It indicates a 2.5 amp slow blow fuse for 120 volt for us in the USA. I would start with that and increase until it doesn't blow on a cold startup. 2.5 amp slow blow is probably good enough. I used a 3 amp cause I didn't have a 2.5 amp slow blow fuse.
Rush
A quick calculation for fuse is 400VA/120V = 3.3A (for 120VAC), so 3A or 3.5A. The CL60 in the primary circuit will reduce the inrush current a bit so the 3A is a good place to start. Or as Rush says, try a 2.5A.
However the blowing of the four 1.1R 3W pi resistors is a huge concern. The equivalent resistance is .275R 12W so many amperes of current is required to blow them. So first, a thorough check of your power supply needs to be done (no amplifier boards attached). I would start by disconnecting the transformer and checking each secondary for the correct voltage. Then connect the diode bridges and check for correct DC voltage from both bridges. Then check your PS board. Since the pi resistors are blown, remove them and then check your capacitors to make sure none are shorted. Replace the pi resistors, but use your meter and confirm their resistance values before installing.
For testing, definitely use a light bulb current limiting tester in case of short. Test the repaired board with no load. If it passes then you can test the amplifier circuit, but with only one channel connected. At this point it is probably best to test without the VFETs in place. Test each channel this way. If okay, then test one channel with VFETs in place. Then test the other channel only with VFETs in place.
However the blowing of the four 1.1R 3W pi resistors is a huge concern. The equivalent resistance is .275R 12W so many amperes of current is required to blow them. So first, a thorough check of your power supply needs to be done (no amplifier boards attached). I would start by disconnecting the transformer and checking each secondary for the correct voltage. Then connect the diode bridges and check for correct DC voltage from both bridges. Then check your PS board. Since the pi resistors are blown, remove them and then check your capacitors to make sure none are shorted. Replace the pi resistors, but use your meter and confirm their resistance values before installing.
For testing, definitely use a light bulb current limiting tester in case of short. Test the repaired board with no load. If it passes then you can test the amplifier circuit, but with only one channel connected. At this point it is probably best to test without the VFETs in place. Test each channel this way. If okay, then test one channel with VFETs in place. Then test the other channel only with VFETs in place.
Last edited:
Ben, I think he had only one 1.1 ohm resistor per rail (4 rails). In-rush current may have been enough to blow them.
Hard to say without his schematic and pictures.
Rush
Thanks a lot Ben. Good plan. So far, I've got until diode bridge output ok. For both polarities.
I had 4x 1.1 ohm resistors in parallel per PSU polarity. So, total of 8 for the amp. This is the diyaudio universal PSU v3 board.
I used a light bulb tester for the first few power ons.
Per rodeodave, got 100w/72w actual incandescent for load testing as well.
I also got a 3A fuse. 2A and 3A were easily available locally.
I had 4x 1.1 ohm resistors in parallel per PSU polarity. So, total of 8 for the amp. This is the diyaudio universal PSU v3 board.
I used a light bulb tester for the first few power ons.
Per rodeodave, got 100w/72w actual incandescent for load testing as well.
I also got a 3A fuse. 2A and 3A were easily available locally.
