Recently built a transistor amp. I'm using a 4ohm load. Its a sub amplifier.
Tried using 2pairs of 2sc5200(Quasi Complimentary).
The amp works great at volume set to 3O'clock for about 30 min.
Heatsink gets hot, but I can still touch it. Then suddenly I loose one of the 2sc5200's. They are all Toshibas and have the same batch numbers.
Am I missing something? How many pairs of these do I need for 4Ohm load and +/- 30V supplies(fully loaded) I measure into 4ohm resistive load, 112W before clipping.
I'm now thinking of using Mjl21194's. But I want to be sure before they go up in smoke too. I'm sure the rest of the circuit is working. I have my Vbe multiplier glued to one of the transistors. Biasing a bit cold.🙁🙁
Tried using 2pairs of 2sc5200(Quasi Complimentary).
The amp works great at volume set to 3O'clock for about 30 min.
Heatsink gets hot, but I can still touch it. Then suddenly I loose one of the 2sc5200's. They are all Toshibas and have the same batch numbers.
Am I missing something? How many pairs of these do I need for 4Ohm load and +/- 30V supplies(fully loaded) I measure into 4ohm resistive load, 112W before clipping.
I'm now thinking of using Mjl21194's. But I want to be sure before they go up in smoke too. I'm sure the rest of the circuit is working. I have my Vbe multiplier glued to one of the transistors. Biasing a bit cold.🙁🙁
How many pairs
2+2 on a +/-30v rail is fine for 4 ohms. Please post your schematic.
If you are using a double darlington the bias spreader should be set for about 2.2v maximum with typical 0.1- 0.22R emitter resistors.
If the output stage is a triple then set at about 3.3v
Regards
Steve Mantz
Zed Audio
PS How is Pretoria? I have not been there since I left SA 29 years ago.
2+2 on a +/-30v rail is fine for 4 ohms. Please post your schematic.
If you are using a double darlington the bias spreader should be set for about 2.2v maximum with typical 0.1- 0.22R emitter resistors.
If the output stage is a triple then set at about 3.3v
Regards
Steve Mantz
Zed Audio
PS How is Pretoria? I have not been there since I left SA 29 years ago.
Monitor current through their ballast resistors. WIth voltage and current, then see the data sheet for those transistors, and look at the "SOA" chart - Safe Operating Area. see if you are violating its limits. A 16A 200v transistor can do 16 amps or 200 volts, but not both at the same time.
Also, look into your cooling. Is the heatsink adequate and can it cool itself enough?
Monitor power rail currents on a scope. Any spiking as signal flows? You say bias is cold. You say that from what measurements? If for some reason bias is not tracking, it may be draw gets excessive when the amp is pushed.
Also, look into your cooling. Is the heatsink adequate and can it cool itself enough?
Monitor power rail currents on a scope. Any spiking as signal flows? You say bias is cold. You say that from what measurements? If for some reason bias is not tracking, it may be draw gets excessive when the amp is pushed.
Greetings from another Ex SA..... but I do miss my Castles and boerewors.
I recommend three pairs of toshibas for 30v rails, 4 ohm loads especially if amp is going to be driven hard and long periods but make sure the amp is not oscillating when driven hard or when it clips. Speakers of 4 ohms can present a impedance as low as 1 ohm at certain frequencies and the reason I recommend 3 pairs, it could also be reason oscillation occurs and or during clipping.
2 pairs of 21194s are however enough but there will be detoriation of THD specs ect, paper spec not necessarily sound wise.
With a schematic we could better advise.
112W into 4ohm 60degree phase angle with Tc=80degC looks OK.
But how can you get 29.93Vpk at the speaker, when the supply rails are at +-30Vdc?
My model predicts ~+-39Vdc for the supply.
Then the 2pair run into difficulties trying to push 12Apk when @ 20Vce.
Reducing Tc<58degC solves this problem.
But how can you get 29.93Vpk at the speaker, when the supply rails are at +-30Vdc?
My model predicts ~+-39Vdc for the supply.
Then the 2pair run into difficulties trying to push 12Apk when @ 20Vce.
Reducing Tc<58degC solves this problem.
For a sub amp MJL21194 may be better. One thing you need to check is that your parts are genuine - 2SC5200 are commonly faked
the original specified pairs will work if you can ensure they never get hot.
Hi Didi,
you can also try the MJL3281A, I found them to be pretty hardy and robust devices. Mantech sells them for around R60 for small quantities.
If you are using a typical LC filter on the speaker and the bass amp runs into clipping (which is very easy at ±30VDC supply at a low frequency), the capacitor essentially looks like a momentary short circuit and the transistors are at risk.
you can also try the MJL3281A, I found them to be pretty hardy and robust devices. Mantech sells them for around R60 for small quantities.
If you are using a typical LC filter on the speaker and the bass amp runs into clipping (which is very easy at ±30VDC supply at a low frequency), the capacitor essentially looks like a momentary short circuit and the transistors are at risk.
A general rule of thumb for me is 1 pair for every 40 to 50W of output. So for a 100W amp I would use 2 pairs and for a 200 water at least 4 pairs. If you expect to drive nasty reactive loads, err on the side of caution and add an extra pair once you get above 100W.
Do you have a p,ot of you speakers impedance and phase? Might be the answer to your problem lies in that data.
Do you have a p,ot of you speakers impedance and phase? Might be the answer to your problem lies in that data.
Hi Nico.
Is jy Afrikaans? Ek is, maar dit maak ni saak ni. Engels: dan verstaan almal HEHE.
Anyways, Hugh told me about you. Great to know there are some serious audio minded GURUS such as yourself in SA. I will check with Mantech and order some. I luckily have an account with them. So do you suggest i get rid of Zobel?
Hi didi, I am totally bilingual.
To be honest with you I am not convinced by haphazardly using the proverbial 10 ohms in series with 100nF. The purpose of the Zobel is to make a load appear resistive in stead of reactive.
I do not believe that all loudspeakers have the exact same characteristic, thus a standard Zobel network as used by most in audio may not have the desired result.
Mind you, one can always detect that an amp is oscillating by a smoking 10 ohm Zobel resistor.
There may be another reason for the so called Zobel, to act as a low pass filter reducing the susceptibility of the amp to conducted RFI via the speaker cord.
To be honest with you I am not convinced by haphazardly using the proverbial 10 ohms in series with 100nF. The purpose of the Zobel is to make a load appear resistive in stead of reactive.
I do not believe that all loudspeakers have the exact same characteristic, thus a standard Zobel network as used by most in audio may not have the desired result.
Mind you, one can always detect that an amp is oscillating by a smoking 10 ohm Zobel resistor.
There may be another reason for the so called Zobel, to act as a low pass filter reducing the susceptibility of the amp to conducted RFI via the speaker cord.
It's a temperature problem, you may still be able to touch the sinks but the transistor dies are probably much too hot. Improve the thermal coupling to the 'sinks and you may also need to increase the sink size.
Regards
Henry
Regards
Henry
- Status
- Not open for further replies.