Hi all,
I completed my first amp a few months ago. It was more a test than really something to listen to. It used the 9240/240 Mosfets.
Now I want to make a more serious amp, but still keep the cost low and the circuitry simple. I've found a very easy schematic with not much components. I'm going to make a few of them to sell to my friends.
I'm going to add somethings myself; thermal compensation, DC protection, inrush current limiter,..
This is the website where I found it 100 W 8 ohms MOSFET amplifier with IRF9540 - IRF540 | Electronics Circuits Diagrams
I know some of you don't like V-Fets and schematics from that kind of website but what do you think of the circuit? Any major drawbacks?
Thanks
I completed my first amp a few months ago. It was more a test than really something to listen to. It used the 9240/240 Mosfets.
Now I want to make a more serious amp, but still keep the cost low and the circuitry simple. I've found a very easy schematic with not much components. I'm going to make a few of them to sell to my friends.
I'm going to add somethings myself; thermal compensation, DC protection, inrush current limiter,..
This is the website where I found it 100 W 8 ohms MOSFET amplifier with IRF9540 - IRF540 | Electronics Circuits Diagrams
I know some of you don't like V-Fets and schematics from that kind of website but what do you think of the circuit? Any major drawbacks?
Thanks
Hi all,
I completed my first amp a few months ago. It was more a test than really something to listen to. It used the 9240/240 Mosfets.
Now I want to make a more serious amp, but still keep the cost low and the circuitry simple. I've found a very easy schematic with not much components. I'm going to make a few of them to sell to my friends.
I'm going to add somethings myself; thermal compensation, DC protection, inrush current limiter,..
This is the website where I found it 100 W 8 ohms MOSFET amplifier with IRF9540 - IRF540 | Electronics Circuits Diagrams
I know some of you don't like V-Fets and schematics from that kind of website but what do you think of the circuit? Any major drawbacks?
Thanks
An important disadvantage is the use of apparently but not really "true complementary" output stage and thus too much differences between the pos. and neg. half.
Avoid additional typical quasi complementary designs like that one from post #166 about
http://www.diyaudio.com/forums/solid-state/43331-power-amp-under-development-17.html#post601021
because the clipping behaviour isn't good.
I would prefer this circuit:
http://www.amplimos.it/images/N-CH1.JPG
ask Mr. Antony Holton about
http://www.aussieamplifiers.com/
and check out this thread:
http://www.diyaudio.com/forums/soli...better-audio-non-complements-audio-power.html
Last edited:
Conclusion; this circuit is not going to explode into my face when I turn it on (unless I did something stupid)
Thanks for the info tiefbassuebertr! But I want to use this design, I've already got a +-35V supply so I'm going to order the IRF540/9540, other components still here from 9240/240 project. Hopefully it works, then I'm going to order a bunch of PCB's and make my friends pockets lighter
Thanks for the info tiefbassuebertr! But I want to use this design.
Thanks for the info tiefbassuebertr! But I want to use this design, I've already got a +-35V supply so I'm going to order the IRF540/9540, other components still here from 9240/240 project. Hopefully it works, then I'm going to order a bunch of PCB's and make my friends pockets lighter
Thanks for the info tiefbassuebertr! But I want to use this design.
Yes irf540 is the cheapest Rp5000 here (~$0.5), but without classH or tracking rail, it will need big heatsink. Their TO220 case not suitable for dissipating power more.
Btw total cost of my 200W into 4 ohm with irf540/9540 is around $50 (large 5A 32V transformer, 2x10kuF, no case), but I am not to sell it here.
Btw total cost of my 200W into 4 ohm with irf540/9540 is around $50 (large 5A 32V transformer, 2x10kuF, no case), but I am not to sell it here
.I think I've found a more reliable design
I've simulated this and it works fine, 200mV is near clipping.
Only problem that I have (it's a simulation) is that I cant adjust bias current and when I put a dummy load on it it draws 37A. Probably a problem with Multisim because the scope shows a perfect sinewave.
I've also simulated the design in the first post, Multisim shows exploding caps and resistors ^^
source: http://www.redcircuits.com/Page2.htm
I've simulated this and it works fine, 200mV is near clipping.
Only problem that I have (it's a simulation) is that I cant adjust bias current and when I put a dummy load on it it draws 37A. Probably a problem with Multisim because the scope shows a perfect sinewave.
I've also simulated the design in the first post, Multisim shows exploding caps and resistors ^^
source: http://www.redcircuits.com/Page2.htm
Simulation
Any ideas what's wrong? Or just a problem with te software?
Input signal is 200mV, 1kHz.
An externally hosted image should be here but it was not working when we last tested it.
Any ideas what's wrong? Or just a problem with te software?
Input signal is 200mV, 1kHz.
Hi
I suspect a real world version of this circuit would self destruct with a fast signal transient. You need to AC couple the gates of the N-ch and P-ch outputs. Without significant bias, >100mA, crossover distortion products will be high. With adequate bias, these devices need thermal compensation. It is pointless to dredge through simulation problems if you know there will be problems with the 'real' circuit, that is if your intent is to build an actual amp. OTOH, Learning simulation parameters and experience is valuable info.
As for the circuit in question, the input stage Z out may be too high to drive the VAS. Try a VAS buffer. What is the bias of the output devices?
I suspect a real world version of this circuit would self destruct with a fast signal transient. You need to AC couple the gates of the N-ch and P-ch outputs. Without significant bias, >100mA, crossover distortion products will be high. With adequate bias, these devices need thermal compensation. It is pointless to dredge through simulation problems if you know there will be problems with the 'real' circuit, that is if your intent is to build an actual amp. OTOH, Learning simulation parameters and experience is valuable info.
As for the circuit in question, the input stage Z out may be too high to drive the VAS. Try a VAS buffer.
What is the bias of the output devices?
Last edited:
Would be around 100-150mA, this design has been built before and is sold as a kit, so I think it should be okay to build this
Facts and Figures - RED
Facts and Figures - RED
What do you mean and how is this done?
Every transistor, whether BJT or mosfet, requires energy to turn off as well as turn on. If, during a large fast transient, one transistor turns on before the other turns off then cross conduction will destroy them. A cap between the gates, before the gate stopper, will provide AC coupling. The cap, ~10-100uf, provides a low impedance 'turn off' source for AC by coupling it to the oposite polarity driver. This cap would be in paralell with the bias setting, or Vgs multiplier if you choose to use one. I am only giving suggestions, I see others may have had luck with this circuit as is, but adding source ballast resistors and a Vgs multiplier would improve bias stability and improve reliability. Circuits originally designed by others can always be improved.
Another approach, which I personally prefer, is to use class A bias totem pole drive for each gate. I would certainly add an emitter follower driver stage. This takes a lot of the large non-linear capacitive loading from the mosfets off of the VAS.
Another approach, which I personally prefer, is to use class A bias totem pole drive for each gate. I would certainly add an emitter follower driver stage. This takes a lot of the large non-linear capacitive loading from the mosfets off of the VAS.
Last edited:
I think I've found a more reliable design
I'm not liking how the bias is set on the output devices. Using a resistor may work at one temperature, but you run the risk of having the output run away on you. Thermally that is... I'd use a bias splitter that tracks the Vgs of the output devices.
~Tom
With the input grounded, do you get reasonable DC operation points? I.e. is the quiescent current through the output devices reasonable? If not, tweak the resistance of the pot.
~Tom
For the N-th time, this is an idiot's rework of a simple alteral MOSFET schematic with IRF parts stuck in without knowledge or thinking. It appear on here at least once a year with alarming regularity. I'm sure someone will also assure us that they have tried it and it works (as in, did not blow up in their faces immediately).
It IS possible to make it that simple, as for performance, that's another story.
1) Add thermal compensation - a Vgs multiplier will work best here, it's even possible to dispense with the source resitors in that case.
2) Add gate protection. This amp will eventually destruct if any unfiltered HF is given to it.
3) AC couple the gates of ht emosfets (already mentioned in the thread)
4) Add gate stoppers
5) DO NOT use IRF540 and 9540 as complementary pair because they are NOT electrically complementary, but metalurgically. This should be obvious from the different maximum currents mentioned right at the beginning of the datasheet (but who reads those, anyway, doh?!). Use IRF640 and IRF9540, regardless of different maximum Vgs - as long as max Vgs is below the voltage between +V and -V, it is an IRRELEVANT PARAMETER FOR CHOOSING COMPLEMENTS!!!!
With all that and perhaps some attention to detail like some degeneration in the input differential, you may actually get a simple amp that performs reasonably well at least...
It IS possible to make it that simple, as for performance, that's another story.
1) Add thermal compensation - a Vgs multiplier will work best here, it's even possible to dispense with the source resitors in that case.
2) Add gate protection. This amp will eventually destruct if any unfiltered HF is given to it.
3) AC couple the gates of ht emosfets (already mentioned in the thread)
4) Add gate stoppers
5) DO NOT use IRF540 and 9540 as complementary pair because they are NOT electrically complementary, but metalurgically. This should be obvious from the different maximum currents mentioned right at the beginning of the datasheet (but who reads those, anyway, doh?!). Use IRF640 and IRF9540, regardless of different maximum Vgs - as long as max Vgs is below the voltage between +V and -V, it is an IRRELEVANT PARAMETER FOR CHOOSING COMPLEMENTS!!!!
With all that and perhaps some attention to detail like some degeneration in the input differential, you may actually get a simple amp that performs reasonably well at least...
I havent found this a problem with my irf240/9240 designs.
Any differences are taken care of by bias and feedback.
IRF540 and IRF9540 differ significantly in Gm. I believe this is what ilimzn is referring too. IRF640 is a much better complement in this respect to IRF9540 than IRF540. This is important in reducing crossover and other distortion components created by the missmatch in Cin vs Gm vs Vgs.
But then there are a few people who seem to have 'magic feedback loops' with infinate bandwidth and zero phase margin that can correct all the extra distortion created by not using the proper components, layout, and circuit.
But then there are a few people who seem to have 'magic feedback loops' with infinate bandwidth and zero phase margin that can correct all the extra distortion created by not using the proper components, layout, and circuit.
...Which is one more thing left to feedback to magically correct. If I remember right, the point was to create as good an amplifier as you can before applying feedback...
IRF(P)240 and 9240 differ in Gm about 2:1, unless you specially select them for very good 9240s and very bad 240s. This means the crossover distortion will have it's own even order distortion because the crossover region is not symetrical -> 4th harmonic and above will be created. 240 and 9140 are much better complements. That being said, devices differ not only because of normal tolerances, but also because of process variations by different manufacturers of nominally the same device. Also, the actual difference in Gm for P and N parts that are metalurgical complements depends on Vdsmax, for lower voltages N types signifficantly overtake the P parts if HEXFETs are in question. 9140 and 240 are better complements than 9240 and 340, the latter pair being higher voltage parts. Granted, IRF(P)340 is not so easy to obtain and not regulairly stocked, mostly because the distributors are also on the '9240 is the complement of 240' bandwagon.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.