Hello, i have some mosfet's laying around, and i want to build my first amp with mosfets ( I ve build many BJT power amplifiers) but never with mosfets
I found this on our friend Google.
With +-24v i think it can put around 50w into 4ohm
Here's the schematic, what do you think?
I only have N-channel mosfets...so I would like to build a quasi.
Cheers
I found this on our friend Google.
With +-24v i think it can put around 50w into 4ohm
Here's the schematic, what do you think?
I only have N-channel mosfets...so I would like to build a quasi.
Cheers
Attachments
Last edited:
First quasi comp fet amp I've seen.
Looks easy enough, though there are kind of a lot of parts for point to point.
Must be a reason nobody does this.
You give up on the 2n3055 amp? I got 21 vac out of my single output pair transistor AX6 last night finally, on Rihanna Shut Up & Drive. Thats 55 watts/channel @ 8 ohms, and those bass notes in that song aren't very short, either. 70 v single voltage power supply 6.5 amp transformer, single 3300 uf cap both channels.
I used a CD. I'm afraid to use the turntable that loud, the last time I tried it the suspension started vibrating with the bass notes.
http://www.diyaudio.com/forums/solid-state/236256-retro-amp-50w-single-supply-20.html
Looks easy enough, though there are kind of a lot of parts for point to point.
Must be a reason nobody does this.
You give up on the 2n3055 amp? I got 21 vac out of my single output pair transistor AX6 last night finally, on Rihanna Shut Up & Drive. Thats 55 watts/channel @ 8 ohms, and those bass notes in that song aren't very short, either. 70 v single voltage power supply 6.5 amp transformer, single 3300 uf cap both channels.
I used a CD. I'm afraid to use the turntable that loud, the last time I tried it the suspension started vibrating with the bass notes.
http://www.diyaudio.com/forums/solid-state/236256-retro-amp-50w-single-supply-20.html
Last edited:
Can you post pictures with exactly what pcb did u use fpr the ax6, I've read the post and i want to build it.
Also what transistors did you used?. Your transformer is 6amp total? 3 amp per channel or?. Thanks again
Also what transistors did you used?. Your transformer is 6amp total? 3 amp per channel or?. Thanks again
I built the AX6 point to point. The picture is on the link in post 2, post # 212. I used Nema CE board .062"x3.5"x5.5" from mcmaster.com sawed from a 12"x24" blank. I used kynar insulated 28 ga wire for the front end stuff (the white ones) and 24 ga teflon insulated wire from driver to output transistors and out (the blue and orange wires). I like kynar & teflon wires, they don't burn when I hit them with the soldering iron. Red & black wires from the regulator board to the storage capacitors and to fuse is 22 ga. Rail fuse is 10 amps.
I was using NTE60 output transistors, which have the same specs roughly as On Semi MJ15003. I use Fairchild TIP41C/42C as drivers, although with 6 mhz Ft they don't do tinkly bells as well as the original dynaco ST120 channel using 1970 RCA selected TO5 drivers. VAS is a GE D44R2 with 20 mhz Ft. Input transistor is an On Semi 2N5401 with incredible gain. That allowed me to make input impedance >100k for compatibility with my PAS2 preamp. If I build AX6 again I'll use MJE15031/32 as drivers, they are faster.
The AX6 side is quieter idle than the original dynaco ST120 side with 1970 transistors .
Transformer was dynaco ST120 original 6.5 amps 80v average open circuit regulated to 70vdc by 5 parallel TIP142 with .22 emitter resistors and a 6x12v 1.2 W zener stack on the base. 900 ohm zener current limiter resistor.
Apex's board looks artwork competent but I don't like importing things like boards from ***** and I don't have any layout software that works on my Linux operating system. His artwork wouldn't fit in my ST120 case anyway, it is too tall or wide. My output transistors are on the original dynaco heat sinks, supplemented by PCAT supply fans. 2.65 amps/channel for 2 seconds last night on a 4 minute song, no audible distortion. Note the single supply design with output capacitor, a 3300 uf per channel with 11 windings around the capacitor as an RF filter on the speaker line. The case has a 1 kohm 3 watt resistor series .47 uf capacitor for a zobel network on the back of the speaker banana jacks, post speaker capacitor.
I think your post 1980 2n3055's would work fine for brief periods at 2.65 amps or 55 W into 8 ohm, Thompson's datasheet shows a second breakdown current limit of 2.87 amps at 40 v. The NTE60 OT I have SOA limit is 5 amps at 50v for 1 second. If you still want to use 2 ohm speaker, use 3 parallel 2n3055 pairs on your heat sink and fly the driver wires over like I did. Use .5 ohm emitter resistors. the TIP41C/42C drivers would be fine for a triple OT to a woofer, have 8 amp drive, and only cost $.38 each. I used a heat sink on the drivers. The TIP41C/42C are just a little slow for the highest of tinkly bell tracks through the tweeter. It was the Martin Denny Hawaii Album where i noticed the Dynaco ST120 with djoffe bias board side sounding better.
Have fun.
I was using NTE60 output transistors, which have the same specs roughly as On Semi MJ15003. I use Fairchild TIP41C/42C as drivers, although with 6 mhz Ft they don't do tinkly bells as well as the original dynaco ST120 channel using 1970 RCA selected TO5 drivers. VAS is a GE D44R2 with 20 mhz Ft. Input transistor is an On Semi 2N5401 with incredible gain. That allowed me to make input impedance >100k for compatibility with my PAS2 preamp. If I build AX6 again I'll use MJE15031/32 as drivers, they are faster.
The AX6 side is quieter idle than the original dynaco ST120 side with 1970 transistors .
Transformer was dynaco ST120 original 6.5 amps 80v average open circuit regulated to 70vdc by 5 parallel TIP142 with .22 emitter resistors and a 6x12v 1.2 W zener stack on the base. 900 ohm zener current limiter resistor.
Apex's board looks artwork competent but I don't like importing things like boards from ***** and I don't have any layout software that works on my Linux operating system. His artwork wouldn't fit in my ST120 case anyway, it is too tall or wide. My output transistors are on the original dynaco heat sinks, supplemented by PCAT supply fans. 2.65 amps/channel for 2 seconds last night on a 4 minute song, no audible distortion. Note the single supply design with output capacitor, a 3300 uf per channel with 11 windings around the capacitor as an RF filter on the speaker line. The case has a 1 kohm 3 watt resistor series .47 uf capacitor for a zobel network on the back of the speaker banana jacks, post speaker capacitor.
I think your post 1980 2n3055's would work fine for brief periods at 2.65 amps or 55 W into 8 ohm, Thompson's datasheet shows a second breakdown current limit of 2.87 amps at 40 v. The NTE60 OT I have SOA limit is 5 amps at 50v for 1 second. If you still want to use 2 ohm speaker, use 3 parallel 2n3055 pairs on your heat sink and fly the driver wires over like I did. Use .5 ohm emitter resistors. the TIP41C/42C drivers would be fine for a triple OT to a woofer, have 8 amp drive, and only cost $.38 each. I used a heat sink on the drivers. The TIP41C/42C are just a little slow for the highest of tinkly bell tracks through the tweeter. It was the Martin Denny Hawaii Album where i noticed the Dynaco ST120 with djoffe bias board side sounding better.
Have fun.
Last edited:
There is this very well designed PCB.
Attachments
Datasheetcatalog.com has got a better datatable, but, there is no SOA rating or thermal resistance rating. They have a slightly better hfe minimum 15 than the RCA 1970 2n3055 which was 5. Ft 2 mhz so make sure you use the 10 ohm base stopper resistors, which AX6 has. No telling about KD503 current toughness except they are too weird to be counterfeits. Wattage is 150, Andrew says in general don't use more than 1/3 of that for reliability.
The datasheet was in German or something similar. If that was GDR, east Ger, then in general the products made for the military were first class, the products made for consumers were ****.
Tip on AX6 or any single supply amp. The gain of the input and VAS transistors you actually use determines where the middle point (input of output cap) ends up. You want the middle to be half the supply rail so that you have maximum headroom to avoid clipping. So one adjusts the pull up and pulldown resistors on the base of the input transistor to get the output up where you need it. You'll notice on my post #212 of AX6 thread, I describe where I ended up on those resistors. Way higher in value than designer used, because I wanted higher input impedance than the design has, and my input transistor was very high gain.
Last edited:
I have done a couple of quasi mosfet amps.
One was lateral mosfets and the other vertical mosfets.
They sound very good.
One was lateral mosfets and the other vertical mosfets.
They sound very good.
Here is a mosfet quasi design.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
I can't see where any one else has given advice on your input Long Tail Pair.
Here's mine.
Look at the current through R13. = ~ [Vgrnled -0.6V] / 68r = 18mA
Much of that will flow through R11 giving a Vdrop ~18*68 = 1200mV
Add on Vbe Q4 to find the voltage across R3. = 1.2 + 0.6 = 1.8V
I r3 = 1.8/1200r = 1.5mA
Now adjust R6 to ensure that the Ie of both Q1 and Q2 are identical.
Remove Vr1 and replace with two ~100r, that are matched to better than 0.1%
Adjust R6 so that Vdrop on these emitter resistors are identical.
Now you have an balanced LTP.
Next ensure the resistance (@ dc) of both the LTP bases are identical.
at the moment you have 1k||10k on Q2b = 909r and 1k+47k on Q1b = 48k
This will give a big output offset.
To help with this choose Q1 and Q2 to have close matching of Vbe and hFE when measured at the same operating current you want to pass.
Now move on the the whole amplifier.
Add passive Low Pass and High Pass filters to the input to define the pass band you need for your duty.
Add a DC blocking capacitor in series with R8 to reduce the gain of the amplifier to ~1times at DC.
Change the output Zobel to Thiele Network
try the following values. R20 = 5r to 10r
C1 = 47nF to 150nF
Add output inductor from 0u5H to 2uH
Add parallel damping resistor, try 2r to 5r
Add supply rail decoupling.
try 1mF at Q9 drain and same at R19.
try 220nF at Q7 collector and at R11 and at R15 and at R13
try 220uF at R4 and at D1.
Add filtering resistance between D1 and R13, try 330r
Split R7 into two equal ~ 11k each. Add 100uF from junction to -ve rail.
Join Input signal return to R1. Add a resistor||inverse diodes from R1 to decoupling ground.
Connect C1 to decoupling ground. Join speaker return to decoupling ground. (I do not recommend the use of a generalised GND, use specific links).
Last edited:
- Status
- Not open for further replies.