Tweet
Sometimes I dread SVT4-Pro’s arriving with no output, though hopeful it’s something simple. Not this time. With a couple probes monitoring the output busses of the two amp channels at the Source resistor community between the heat sinks, as I began bringing up the variac, Ch A’s offset climbed Positive immediately, indicating a complete failure of the P-Ch devices and source resistors (all open). Ch B looked nominal, for a far as I advanced the variac (10VAC mains).
The ‘dread’ aspect was thinking I had exhausted all of my previous batched/matched IRFP240’s and IRFP9240 MosFET’s, but, did have two fresh tubes of 20 pcs each, though not yet screened/batched and matched. I pulled the amp apart, set the chassis and rear panel modules aside, leaving the Power Amp PCB assembly in the test bench for scrutiny.
Power Amp PCB Schematics (428xxh0_).pdf
Power Amplifier Pictorials (428XXp9).pdf
The P-Ch devices were all shorted, source resistors all open, gate resistors all open, current sense resistors fried on the P-Ch side, neg driver xstr shorted, neg current limiter xstr shorted. N-Ch devices were measuring bad, though one was obviously shorted. After making the beginning of a parts list, I began pulling parts off the rear heat sink so I could remove the heat sink, as so many of the output stage parts are underneath the bottom relief of the heat sinks.
This isn’t the first time the Ch A side had failed…last time in Sept 2015, and had my notes on that. This time around, the scorching of the sense resistors damaged a few traces & solder pads, leaving a crusty surface on the plated pads. Removing the MosFET’s and gate resistors also did further harm to some solder pads, as careful as I went in the surgery. Only one of the five N-Ch parts failed. The other four measured ok with DMM testing.
Still, I needed to power up the Hameg HM-8042 Curve Tracer (stand-alone unit, XY output coupled to a Tek 7633 Storage Scope/Plug-ins). I have somewhat of a love-hate relationship with that Hameg, as it is very tedious to set up, and changing from N-Ch to P-Ch, you have to go thru all that again. You also need to have some idea of the gate voltage range for the MosFET’s being tested (in this case), or you’ll be chasing your tail forever before finding the right range and begin seeing the results on the screen. Once there, however, it is nice, since you also get digital readout of Vgs, Vd, Id relative to where you’ve set the cursor…being able to scroll thru a range of 5 gate voltages and the range of supply voltage. The Hameg only lets you get to 40V max in the Vd (or Vc), but it does at least yield useful results.
I found I still had 5 P-Ch parts left over from the previous screenings, as well as 8 N-Ch parts. The N-Ch part was close to twice the Id/Vgs value as the rest of the N-Ch stage, which I re-tested. I had enough of the P-Ch parts, though the Id range on the two sets were not equal. In circuit, Vgs on the positive and negative sides aren’t equal anyway (3.86V vs -3.23V)
I carefully put all the fresh parts in, patching circuits back together as needed, wishing for better quality PCB material. Then, I’d need to use my extender cables to couple the pwr xfmr primary wiring to the amp PCB assy and set that on it’s front HS, with the bottom of the PCB facing me for probing. It powered up just fine, thankfully. I got the idle measurements as follows:
Part Vs Ids Vgs Part Vs Ids Vgs
Q110 17.5mV 37.2mA 3.86V Q111 -18.3mV -38.9mA -3.23V
Q112 32.8mV 69.8mA 3.86V Q113 -21.6mV -46.0mA -3.23V
Q114 18.0mV 38.3mA 3.86V Q115 -23.5mV -50.0mA -3.23V
Q116 21.3mV 45.3mA 3.86V Q117 -22.0mV -46.8mA -3.23V
Q118 20.4mV 43.4mA 3.86V Q119 -22.7mV -48.3mA -3.23V
I was concerned about Q112 being so much higher in Ids. But, this is at idle, and NOT under driven load current.
After verifying I had good output signal with no surprises, I wanted to measure the output current of each of the MosFET’s, looking across their respective 0.47 ohm/5W Source resistors. Now, I didn’t stop to measure & record each of them, other than seeing they all were in the 0.465-0.483 ohm range. Close enough. I used a True RMS HP 3467A 4-Ch DMM with shielded probes, com lead connected to the output buss, and moving the plus lead from Source to Source terminals, recording the results. I began with 10V RMS / 8 ohm load, which would be a 1.25A output current total.
Part Vs RMS Is RMS Part Vs RMS Is RMS
Q110 63.0mV 134mA Q111 58.7mV 125mA
Q112 63.0mV 134mA Q113 62.4mV 133mA
Q114 60.9mV 130mA Q115 64.1mV 136mA
Q116 65.1mV 139mA Q117 61.3mV 130mA
Q118 63.6mV 135mA Q119 61.7mV 131mA
Total current 672mA 655mA
10V RMS output into 8 ohms = 1.25A RMS
N-Ch + P-Ch current 672mA + 655mA = 1.33A RMS
Next, I increased the output to 20V RMS into 8 ohms
Part Vs RMS Is RMS Part Vs RMS Is RMS
Q110 134.3mV 286mA Q111 123.8mV 263mA
Q112 121.3mV 258mA Q113 129.3mV 275mA
Q114 130.0mV 277mA Q115 132.3mV 281mA
Q116 135.1mV 287mA Q117 128.0mV 272mA
Q118 133.0mV 283mA Q119 127.0mV 270mA
Total current 1.39A 1.36A
20V RMS output into 8 ohms = 2.5A RMS, 50W
N-Ch + P-Ch current 1.39A + 1.36A = 2.75A RMS
Finally, I increased the output to 40V RMS into 8 ohms
Part Vs RMS Is RMS Part Vs RMS Is RMS
Q110 271mV 577mA Q111 255mV 543mA
Q112 241mV 513mA Q113 262mV 557mA
Q114 266mV 566mA Q115 266mV 566mA
Q116 273mV 581mA Q117 259mV 551mA
Q118 272mV 579mA Q119 256mV 545mA
Total current 2.78A 2.76A
40V RMS output into 8 ohms = 5A RMS, 200W
N-Ch + P-Ch current 2.78A + 2.76A = 5.54A RMS
While the output current per device is half-wave, the readings are no doubt distorted by that, and the sum of the two stages are higher than the sinewave output, it’s showing the current sharing of all the devices aren’t grossly out of whack. I was expecting Q112 to be hogging current with it’s higher Vs curve. In fact, just the opposite as the output current increases under load. Q110, 114, 116 & 118 average to 576mA @ 40V output, with Q112 being only 513mA, 11% low.
With all this seen, I put the amp back together, and felt better seeing my fears of potential failure not being realized.
Back in 2015, I had begun buying the IRFP240’s and IRFP9240’s in bulk, and screening them, sorting per Is/Vgs, and then for Vgs at 83V Vd/50mA Id for both. The nominal supply voltages on the SVT4-Pro is +/- 83VDC. In 2015, the N-Ch and P-Ch matched parts from Ampeg were about $13 ea, $130/ch, $260 per amplifier (both channels). The discrete parts were $1.65 & $2.10 ea, so a lot cheaper to screen them yourself. That was 3 yrs ago…prices have gone up since then both from Ampeg and distributor prices on the MosFET’s.
The ‘dread’ aspect was thinking I had exhausted all of my previous batched/matched IRFP240’s and IRFP9240 MosFET’s, but, did have two fresh tubes of 20 pcs each, though not yet screened/batched and matched. I pulled the amp apart, set the chassis and rear panel modules aside, leaving the Power Amp PCB assembly in the test bench for scrutiny.
Power Amp PCB Schematics (428xxh0_).pdf
Power Amplifier Pictorials (428XXp9).pdf
The P-Ch devices were all shorted, source resistors all open, gate resistors all open, current sense resistors fried on the P-Ch side, neg driver xstr shorted, neg current limiter xstr shorted. N-Ch devices were measuring bad, though one was obviously shorted. After making the beginning of a parts list, I began pulling parts off the rear heat sink so I could remove the heat sink, as so many of the output stage parts are underneath the bottom relief of the heat sinks.
This isn’t the first time the Ch A side had failed…last time in Sept 2015, and had my notes on that. This time around, the scorching of the sense resistors damaged a few traces & solder pads, leaving a crusty surface on the plated pads. Removing the MosFET’s and gate resistors also did further harm to some solder pads, as careful as I went in the surgery. Only one of the five N-Ch parts failed. The other four measured ok with DMM testing.
Still, I needed to power up the Hameg HM-8042 Curve Tracer (stand-alone unit, XY output coupled to a Tek 7633 Storage Scope/Plug-ins). I have somewhat of a love-hate relationship with that Hameg, as it is very tedious to set up, and changing from N-Ch to P-Ch, you have to go thru all that again. You also need to have some idea of the gate voltage range for the MosFET’s being tested (in this case), or you’ll be chasing your tail forever before finding the right range and begin seeing the results on the screen. Once there, however, it is nice, since you also get digital readout of Vgs, Vd, Id relative to where you’ve set the cursor…being able to scroll thru a range of 5 gate voltages and the range of supply voltage. The Hameg only lets you get to 40V max in the Vd (or Vc), but it does at least yield useful results.
I found I still had 5 P-Ch parts left over from the previous screenings, as well as 8 N-Ch parts. The N-Ch part was close to twice the Id/Vgs value as the rest of the N-Ch stage, which I re-tested. I had enough of the P-Ch parts, though the Id range on the two sets were not equal. In circuit, Vgs on the positive and negative sides aren’t equal anyway (3.86V vs -3.23V)
I carefully put all the fresh parts in, patching circuits back together as needed, wishing for better quality PCB material. Then, I’d need to use my extender cables to couple the pwr xfmr primary wiring to the amp PCB assy and set that on it’s front HS, with the bottom of the PCB facing me for probing. It powered up just fine, thankfully. I got the idle measurements as follows:
Part Vs Ids Vgs Part Vs Ids Vgs
Q110 17.5mV 37.2mA 3.86V Q111 -18.3mV -38.9mA -3.23V
Q112 32.8mV 69.8mA 3.86V Q113 -21.6mV -46.0mA -3.23V
Q114 18.0mV 38.3mA 3.86V Q115 -23.5mV -50.0mA -3.23V
Q116 21.3mV 45.3mA 3.86V Q117 -22.0mV -46.8mA -3.23V
Q118 20.4mV 43.4mA 3.86V Q119 -22.7mV -48.3mA -3.23V
I was concerned about Q112 being so much higher in Ids. But, this is at idle, and NOT under driven load current.
After verifying I had good output signal with no surprises, I wanted to measure the output current of each of the MosFET’s, looking across their respective 0.47 ohm/5W Source resistors. Now, I didn’t stop to measure & record each of them, other than seeing they all were in the 0.465-0.483 ohm range. Close enough. I used a True RMS HP 3467A 4-Ch DMM with shielded probes, com lead connected to the output buss, and moving the plus lead from Source to Source terminals, recording the results. I began with 10V RMS / 8 ohm load, which would be a 1.25A output current total.
Part Vs RMS Is RMS Part Vs RMS Is RMS
Q110 63.0mV 134mA Q111 58.7mV 125mA
Q112 63.0mV 134mA Q113 62.4mV 133mA
Q114 60.9mV 130mA Q115 64.1mV 136mA
Q116 65.1mV 139mA Q117 61.3mV 130mA
Q118 63.6mV 135mA Q119 61.7mV 131mA
Total current 672mA 655mA
10V RMS output into 8 ohms = 1.25A RMS
N-Ch + P-Ch current 672mA + 655mA = 1.33A RMS
Next, I increased the output to 20V RMS into 8 ohms
Part Vs RMS Is RMS Part Vs RMS Is RMS
Q110 134.3mV 286mA Q111 123.8mV 263mA
Q112 121.3mV 258mA Q113 129.3mV 275mA
Q114 130.0mV 277mA Q115 132.3mV 281mA
Q116 135.1mV 287mA Q117 128.0mV 272mA
Q118 133.0mV 283mA Q119 127.0mV 270mA
Total current 1.39A 1.36A
20V RMS output into 8 ohms = 2.5A RMS, 50W
N-Ch + P-Ch current 1.39A + 1.36A = 2.75A RMS
Finally, I increased the output to 40V RMS into 8 ohms
Part Vs RMS Is RMS Part Vs RMS Is RMS
Q110 271mV 577mA Q111 255mV 543mA
Q112 241mV 513mA Q113 262mV 557mA
Q114 266mV 566mA Q115 266mV 566mA
Q116 273mV 581mA Q117 259mV 551mA
Q118 272mV 579mA Q119 256mV 545mA
Total current 2.78A 2.76A
40V RMS output into 8 ohms = 5A RMS, 200W
N-Ch + P-Ch current 2.78A + 2.76A = 5.54A RMS
While the output current per device is half-wave, the readings are no doubt distorted by that, and the sum of the two stages are higher than the sinewave output, it’s showing the current sharing of all the devices aren’t grossly out of whack. I was expecting Q112 to be hogging current with it’s higher Vs curve. In fact, just the opposite as the output current increases under load. Q110, 114, 116 & 118 average to 576mA @ 40V output, with Q112 being only 513mA, 11% low.
With all this seen, I put the amp back together, and felt better seeing my fears of potential failure not being realized.
Back in 2015, I had begun buying the IRFP240’s and IRFP9240’s in bulk, and screening them, sorting per Is/Vgs, and then for Vgs at 83V Vd/50mA Id for both. The nominal supply voltages on the SVT4-Pro is +/- 83VDC. In 2015, the N-Ch and P-Ch matched parts from Ampeg were about $13 ea, $130/ch, $260 per amplifier (both channels). The discrete parts were $1.65 & $2.10 ea, so a lot cheaper to screen them yourself. That was 3 yrs ago…prices have gone up since then both from Ampeg and distributor prices on the MosFET’s.