dict.cc dictionary :: Missgeburt :: German-English translation
Thats the english word I was looking for yes.
Miscarriage
If it's a school project, build it, it doesn't matter what it sounds like.
The temperature compensation is the diode D1, and the LED D2, but you would be better off using a VBE multiplier, also known as a amplified diode, if your not sure what it is, try Google-ing it. Fit it to the heatsink between the output devices
If you want to build something to listen to, look elsewhere, I'd recomend the Pass forum, some excellent sounding and easy to build designs, J2 F5 Zen 9, but it will take all your pocket money and more for the PSU and heatsink.
Happy building
The temperature compensation is the diode D1, and the LED D2, but you would be better off using a VBE multiplier, also known as a amplified diode, if your not sure what it is, try Google-ing it. Fit it to the heatsink between the output devices
If you want to build something to listen to, look elsewhere, I'd recomend the Pass forum, some excellent sounding and easy to build designs, J2 F5 Zen 9, but it will take all your pocket money and more for the PSU and heatsink.
Happy building
Try this, compare both schematics, the 2nd one has the Vbe multiplier. Both J L Hood and Dellepiane's original schematics don't have any, soy you could figure it out how to place one for thermal compensation. BTW, you could use a BD139 because you can place it with a screw.
http://www.diyaudio.com/forums/solid-state/167753-linsley-hoods-all-mosfet-power-amp.html
I would recommend mounting the sensing transistor of the VBE-multiplier on top of one of the output devices. This gives a faster response and thus better control.
Although with a conventional BJT VBE-multiplier on MOSFET outputs you would most likely overcompensate the standing current meaning, that it will be greatly reduced (instead of kept constant) when the outputs increase their temperature.
You could place a trim-pot in series with the VBE multiplier circuit - controlling the bias current with this potentiometer and controlling the temperature compensation coefficient by adjusting the VBE multiplier (my guess for the appropriate VBE multiplier multiplication factor would be somewhere around 2 or 3)
This would be the same as 2 or 3 diodes in series with the bias trim-pot. (Like it is already suggested by the schematic). You could glue them on top of one output device. If you have thermal runaway, glue another diode. If you have overcompensation, take one out. If you need to finetune the number of diodes to a non integer value like 2.7 then use a VBE multiplier (with a multiplication factor of 2.7
).
I don't remember if it has a dc offset problem, but I'll measure it these days! A dc offset trimpot it's a great idea.
Ok guys, I've got my PCB right here and guess what, it doens't work.
I've been reading wikipedia and diyaudio all night and I've got the following question;
As an example : my IRFP9240
1. To let this MOSFET work you need to applie NEGATIVE voltage on the GATE wich in my case is -35V or another voltage?
I'm a bit confused right now. I just took the standard footprint out of my PCB design library. At this moment, the drain has got that -35V. The Gate is connected to R13 and the source is connected to R15. You can see wich resistors it are at the first post.
But it's normal that you feed the MOSFET at the source with -35V isn't it...?
Can someone just say whether this is wrong or right so I can go on in my search.
Thanks!
I've been reading wikipedia and diyaudio all night and I've got the following question;
As an example : my IRFP9240
1. To let this MOSFET work you need to applie NEGATIVE voltage on the GATE wich in my case is -35V or another voltage?
I'm a bit confused right now. I just took the standard footprint out of my PCB design library. At this moment, the drain has got that -35V. The Gate is connected to R13 and the source is connected to R15. You can see wich resistors it are at the first post.
But it's normal that you feed the MOSFET at the source with -35V isn't it...?
Can someone just say whether this is wrong or right so I can go on in my search.
Thanks!
Last edited:
IRFP240/9240 are vertical mosfets with very unequal Gm and, not only are they not complements, they require some measure of thermal compensation. These transistors have a Vgs threshold of around 3V so the driving stage cannot fully turn them on during the output voltage peaks. But the most damning thing is the positive temperature coefficient that these divices have. If you are to use these transistors as output devices, you need to have at least a couple hundred mA of bias in order to traverse the steep drop in transconductance that exists at currents below a couple hundred mA, unless you want lots of extra distortion.
This circuit is designed to use lateral mosfets. I suggest using the proper components.....your results will be more pleasing.
This circuit is designed to use lateral mosfets. I suggest using the proper components.....your results will be more pleasing.
Brian,
What an unlucky guy you are. I had it made and it has been up for more than a month. No surprise in either the assembling process or the results. I like the leds, it reminds me whether I had shut the power down.
According to what the original designer said, it wasn't designed to use laterals. But maybe he was inspired by a circuit with such? I did ask him about the Vbe but he said it will overcompensated so I stick with his original schematic.
BTW, I did not use a ammeter to do the calibration. I use a pair of 1/2w 10ohm current limiting resistors and measure the voltage across it. I found this much much easier. If you are using an ammeter, as soon as you see the reading (or the pointer) swings high, you would immediately shut down the power, don't even have to time or courage to adjust the current while it is way over your meter range.
Wonder if you had done that :
Do not start with the 500ohm VR at either end. Turn it at midway and start from there. When I had the project on page 2, I had the current limiter slowly braised when power turned on. It doesn't blow up, just small white fumes and slowly burnt (you cannot even smell it). Pretty sensitive (I think almost all project are like that) at the right point also. If you did not screw up with the pcb and the power mosfet mounting, it shouldn't give you too much of a problem. Don't forget, the mpsa43 and the bc546 have different pin out. Hope you did not mess up here. I did but I found out before I throw the switch. The only problem with this is the 546 looks a bit wierd when soldered to the board.
The 340 are 350 are pretty hot, don't forget to have them "sinked". I use a 4"x7"x1.25" (each channel) heatsink for the mosfet but they still feels pretty hot when I touch them. Well, not burning hot as to draw my hand. Should I say just warm?
Guys, I have got a few of those boards with me. call me if interested : tony_chan_cf@email.com
What an unlucky guy you are. I had it made and it has been up for more than a month. No surprise in either the assembling process or the results. I like the leds, it reminds me whether I had shut the power down.
According to what the original designer said, it wasn't designed to use laterals. But maybe he was inspired by a circuit with such? I did ask him about the Vbe but he said it will overcompensated so I stick with his original schematic.
BTW, I did not use a ammeter to do the calibration. I use a pair of 1/2w 10ohm current limiting resistors and measure the voltage across it. I found this much much easier. If you are using an ammeter, as soon as you see the reading (or the pointer) swings high, you would immediately shut down the power, don't even have to time or courage to adjust the current while it is way over your meter range.
Wonder if you had done that :
Do not start with the 500ohm VR at either end. Turn it at midway and start from there. When I had the project on page 2, I had the current limiter slowly braised when power turned on. It doesn't blow up, just small white fumes and slowly burnt (you cannot even smell it). Pretty sensitive (I think almost all project are like that) at the right point also. If you did not screw up with the pcb and the power mosfet mounting, it shouldn't give you too much of a problem. Don't forget, the mpsa43 and the bc546 have different pin out. Hope you did not mess up here. I did but I found out before I throw the switch. The only problem with this is the 546 looks a bit wierd when soldered to the board.
The 340 are 350 are pretty hot, don't forget to have them "sinked". I use a 4"x7"x1.25" (each channel) heatsink for the mosfet but they still feels pretty hot when I touch them. Well, not burning hot as to draw my hand. Should I say just warm?
Guys, I have got a few of those boards with me. call me if interested : tony_chan_cf@email.com
In source follower configuration? No, look at Vgs threshold in the datasheet, should be between -2V and -4V. Notice that Vgs maximum is +/-20V. Did you remember to include gate source protection Zeners? This is significantly important to prevent shoot through of the oxide layer between the gate and channel because of too much Vgs.
Is the PCB designed for laterals? Perhaps you should compare datasheets. IRFP9240 and 2SK1058 Notice anything different pertaining to the pins labels?
Then those mosfets are probably toasted. Vgs max is +/-20V. I suspect you neglected to include the Zener diode gate source protection.
In the datasheet...
...the symbol shows a body diode from drain to source, which would be reversed bias under normal operating conditions. This is convenient that you don't have to include external flyback diodes. (diodes that arrest any reverse EMF from the speaker that might reverse bias the output devices) If you take that mosfet out of the circuit, measure from drain to source with an Ohm-meter, you should be able to measure that diode if reversed bias on the transistor (for P-ch, + to drain & - to source); infinite reading when forward biasing the transistor (- to drain, + to source). You should see infinite resistance between the gate and either drain or source, either direction. If you measure a resistance between the gate and anything else, that mosfet is 
I suggest re-building the circuit, with everything except the output devices. Temporarily tie the NFB point currently at the output to the cathode of the LED. Then make sure it is working and the gate voltages are correct for those mosfets, Vgs threshold is >2V & <4V. Measure from where the two gates are connected, you should see
2(Vgs) ~4 - 7V. If you do not, then figure out why. If you connect the outputs to the circuit before this voltage is within that range, you risk damaging another set.
...and don't forget the Zener protection.
2(Vgs) ~4 - 7V. If you do not, then figure out why. If you connect the outputs to the circuit before this voltage is within that range, you risk damaging another set.
...and don't forget the Zener protection.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.