10:04 AM 9/21/2022
Hello,
This is a question about the F5 First Watt clone. We built one from parts purchased from DIY. I say "we" built it because a professional colleague is building it for me. It works, but it isn't matching the published voltage on the measurement to adjust bias.
On the first startup we blew the P channel MOSFET (Q4) on one channel because we did not understand the adjustment procedure. I purchased a new kit of MOSFETS and the BJT limiter transistors (Q5, Q6) from DIYAUDIO. We replaced both MOSFETS and the limiter transistors on the blown channel and tried again.
The P channel MOSFET I received in the replacement parts kit is not the same as the original and I questioned that and was assured it was an upgrade. The original was the IRFP 9240 and the upgrade is the IRFP 9141. We did not change the working channel MOSFETs and left the IRFP 9240 in the working channel. This was to keep a spare MOSFET if needed. Why would we need a spare? Smile.
We don't have a variac and used the old trick of houshold light bulbs in series with the AC input to keep things from getting out of hand too quickly. After verifying things were OK, we removed the series light bulbs. Both channels work and have very low DC offset voltage in addition to the ability to play audio.
But, the two channels are very different from each other and also very different from the published bias. The DC supply is +- 23 and is based on the DIY PSU board and published schematic in the DIY build guide. We did not have the PSU board at first and used caps clamped to the chassis floor and a bridge bolted to the front panel.
We can't get the bias past a tenth of the published value of .59 on either channel; we are running at about .059 on both channels. Attached is a table of the values we have. After our first experience of blowing the MOSFET, we watched the voltage drop across the .47 source resistors(R7, R8) as a safety check of power dissipation.
In channel 1 the drop across the source resistors is about 0.700 volts as in the table. In channel 2 the drop across the source resistors is about 0.13 volts. Both channels, as noted above have the same balanced drops across R9 and R11. Channel 1 is running about the right heat to the touch and channel 2 is barely warm to the touch which would be the case with the low drop we see across the source resistors.
We believe that the input JFETS are OK based on voltage measurements and the ability to play audio seems to confirm this. The same input level of audio produces the same output level, as measured by ear, in both channels. P3 was adjusted to the middle mechanically before startup and adjusted for equal drop across R3 and R4.
I am going to ask my colleague, Michael, to describe what happens when he attempts to raise the bias using P1 and P2.
…… Both the channels will only balance at about 0.52 to 0.58 volts. When an attempt is made to go to a high balance point the amount of imbalance stabilizes at about 0.125 volts. At this point adjusting either P1 or P2 drives both voltages in the same direction maintaining the 0.125 volt imbalance. When the higher of the balance voltages reaches the 0.59 volt range the drop across the source resistors gets near or above 1.25 volts which drives the source resistors to 3 watts.……
Thanks for any suggestions. It strikes me that the .7 current across the source resistor is about rith, but why is the bias so low? Attached is table of measured voltages.
Hello,
This is a question about the F5 First Watt clone. We built one from parts purchased from DIY. I say "we" built it because a professional colleague is building it for me. It works, but it isn't matching the published voltage on the measurement to adjust bias.
On the first startup we blew the P channel MOSFET (Q4) on one channel because we did not understand the adjustment procedure. I purchased a new kit of MOSFETS and the BJT limiter transistors (Q5, Q6) from DIYAUDIO. We replaced both MOSFETS and the limiter transistors on the blown channel and tried again.
The P channel MOSFET I received in the replacement parts kit is not the same as the original and I questioned that and was assured it was an upgrade. The original was the IRFP 9240 and the upgrade is the IRFP 9141. We did not change the working channel MOSFETs and left the IRFP 9240 in the working channel. This was to keep a spare MOSFET if needed. Why would we need a spare? Smile.
We don't have a variac and used the old trick of houshold light bulbs in series with the AC input to keep things from getting out of hand too quickly. After verifying things were OK, we removed the series light bulbs. Both channels work and have very low DC offset voltage in addition to the ability to play audio.
But, the two channels are very different from each other and also very different from the published bias. The DC supply is +- 23 and is based on the DIY PSU board and published schematic in the DIY build guide. We did not have the PSU board at first and used caps clamped to the chassis floor and a bridge bolted to the front panel.
We can't get the bias past a tenth of the published value of .59 on either channel; we are running at about .059 on both channels. Attached is a table of the values we have. After our first experience of blowing the MOSFET, we watched the voltage drop across the .47 source resistors(R7, R8) as a safety check of power dissipation.
In channel 1 the drop across the source resistors is about 0.700 volts as in the table. In channel 2 the drop across the source resistors is about 0.13 volts. Both channels, as noted above have the same balanced drops across R9 and R11. Channel 1 is running about the right heat to the touch and channel 2 is barely warm to the touch which would be the case with the low drop we see across the source resistors.
We believe that the input JFETS are OK based on voltage measurements and the ability to play audio seems to confirm this. The same input level of audio produces the same output level, as measured by ear, in both channels. P3 was adjusted to the middle mechanically before startup and adjusted for equal drop across R3 and R4.
I am going to ask my colleague, Michael, to describe what happens when he attempts to raise the bias using P1 and P2.
…… Both the channels will only balance at about 0.52 to 0.58 volts. When an attempt is made to go to a high balance point the amount of imbalance stabilizes at about 0.125 volts. At this point adjusting either P1 or P2 drives both voltages in the same direction maintaining the 0.125 volt imbalance. When the higher of the balance voltages reaches the 0.59 volt range the drop across the source resistors gets near or above 1.25 volts which drives the source resistors to 3 watts.……
Thanks for any suggestions. It strikes me that the .7 current across the source resistor is about rith, but why is the bias so low? Attached is table of measured voltages.
The problem of initial adjustment of a DIYAUDIO First Watt F5 clone has been solved. The adjustment procedure is discussed in the original article and designates the two .47 source resistors as R 11, R12. The design voltage across these resistors is called out in the text and on the schematic in the article.
However, the DIY build guide schematic designates these as R7 and R8. The build guide R11 and R12 are elsewhere in the schematic. So looking for .6 volts here was nearly ten times the dissipation on the source resistors. No wonder we took out a MOSFET in one channel before we watched the dissipation on the source resistors. It was our fault for not reading more carefully, but we did learn something and we actually had to use some of that foundational knowledge we had acquired so long ago. Smile.
However, the DIY build guide schematic designates these as R7 and R8. The build guide R11 and R12 are elsewhere in the schematic. So looking for .6 volts here was nearly ten times the dissipation on the source resistors. No wonder we took out a MOSFET in one channel before we watched the dissipation on the source resistors. It was our fault for not reading more carefully, but we did learn something and we actually had to use some of that foundational knowledge we had acquired so long ago. Smile.
