After three months of hard work I successfully finished my version of the Studio amplifier, designed by Bora Jagodic. Details and schematics about this amp can be found in other topics on this forum, so I won't disscuss them again. What I will talk about is how I made it.
Motivation. I have a pair of great loudspeakers by Sansui. The model is ES209, real rare beaty. They are powerfull speakers with 110W of music power. The main feature of this treasure is a passive radiator which is indeed capable of producing very low frequencies.
I have been using various amplifiers with these speakers and I was not satisfied. Simply, I felt there was more that they could give. I started with integrated amplifier SONY TA-FE370. It is one great amplifier with really good ratio value/quality. However, this amplifier did not fit well with Sansui speakers. The sound was too wide, too detailed and very sharp. I tried the amp with Yamaha speakers as well, but the situation was much better. The conclusion was either to change the speakers or to change the amp.
In the meantime, a friend provided me with two T-Amp amplifiers. I instantly made a powerfull stabilized power supply 2x 12V/4A. I connected T-Amps in biamp configuration and tested them with ES209. The effect was a very nice soft sound, very good quality indeed. Unfortunately, at large power T-Amps created distortion. Its nominal power is 10W, but one amp hardly produced 7W without distortion. Furthermore, the sensitivity of ES209 is around 88dB which required much stronger amplifier. T-Amp generally works well with speaker of >90 dB of senisitivity, so this idea failed as well.
After I explained all this trouble, Bora suggested me the Studio amplifier as an ultimate solution for Sansui ES209.
First steps. As I am DIY fanatic, I decided to make everything from scratch. The first task was to make PCBs and to test capabilities of the amplifier. It was quite easy to make PCBs, since it has no small footprints, such as IC or smd components. It is all discrete, true hole implementation. The next step was to get output transistors. I was lucky enough to find MJL21193 and MJL21194. I got them from a friend who has an electronics shop and he recomended the pair as very good for audio applications. As pre-output stage drivers I used MJE340 and MJE350.
Heatsink. Roaming over flee market I bumped on smuggler who was selling a full box of various heatsinks for a funny price of 20 euro. Of course, I did not miss the chance and bought them all. The box contained various sizes of unused heatsinks, brand new, aluminium. I selected two largest ones and mounted them on the transistors. Misfortune in the decision showed up later when I realized that the orientation of ribs is not fitting the box. I will explain this problem in details in few lines below.
Testing. After the first two boards were assembled I attached a regulater power supply +- 35V/1A. Few minutes were required to tune the quicient currents as design suggested. Then I attached PC CD player and SAL speakers, oscilloscope and a tone generator. Unfortunately, I realized (and heard) that the upper part of a positive semi period of a sinus signal has "hair", i.e. oscillations. On the speakers it sounded as if forest hums due to wind. I was very unhappy.
Finding the bug. Few days after agony and disapointment, I found a guy on a forum who suggested me to change all transistors. This was not making me happy, but I started the tedious work. Luckily for me, the first two transistors I changed were the input ones, BC546, situated in differential pair of the input stage. Instead of BC546 I put BC182 (the ones I had at hand) I turned the power back on and voila! All worked well! I left BC182, following the golden engineering rule "if it works- do not change it". I do not know what was causing the oscillations, but faulty transistors were probably the reason (later on I discarded 50 pcs of those- I did not want to see them again).
Box. I had the opportunity to use a workshop with various tools for metal handling. I asked the owner (who happened to be my father in law, btw.) for the possibility to make one nice metal box. The man was very happy, since he was very bored being in pension. The guy came up the very next day with a finished box outline, shaped on a aluminuim press. The thickness of the material he used was 1.125 mm. I was amazed how fast he made the skeleton. After this point I knew there was no return and that the amp had to be finished.
I decided to spend the whole week in the workshop, drilling, cutting and bending metal. I had never realized before that there was so much work in mechanics for one simple box.
The box was designed in such way that there are five fixed edges which are not moveable. They are welded to each other. The top cover is the only one designed to be able to open. This approach was telling me I would have trouble with mouting the elements inside the box. A logical decision was to make a metal plate which will hold all the electronics and that plate would be fixed for the bottom of the box after the amp is finished. It later showed to be a right decision.
Implementation. I got back to my working desk and started fixing heatsinks and other parts. I used silver wire 2.2mm^2 with teflon insulation. I spent four meters of the wire. I also used KCUG japanese capacitors on critical places. Also, in the filter section I put 4x10,000uF which are bypassed with 330nF/250V Siemens MKP capacitors. The transformer I used is custom made 2x30V AC, 300VA torus. Such transformer required soft start circuit.
For wiring AC power lines, I used a simple copper wire of 2.5mm^2. I also put a fuse f 1.6A (fast) on the primary of the transformer. Finally, the switch used is cutting both AC wires at the same time, thus resolving problems caused by large inductance of the primary.
Protection. I was concerned about my Sansui speakers, since the amplifier was supplied by +-43V DC and the transformer was capable of giving 5A per channel. This is a real danger for speakers. A logical conclusion was that a protection is a must to make.
Since I work with microcontrollers as a professional, I was immediately having an idea of using a PIC to control the soft start circuit and to make protection for speakers. However, I recall reading that PIC and other microcontrollers are not desirable in audio circuits especially on signal path. I was trying to respect this attitude and I designed a protection circuit which is not requiring a direct connection of the PIC to audio signal. The bacics of the protection are taken from this web page: http://sound.westhost.com/project33.htm
Variation of the circuit is made in such way that in collector of Q3 I put an opto-coupler which is then connected to PIC. In this way I separated PIC from audio signal path. I also decided to use a separate power supply for PIC and for the purpose I used a 220/12V transformer.
The protection circuit drives three relays. The first one is Schrak relay for the soft start circuit. It is turned off after 1 second. The other two relays are output relays for speakers. They turn on after 10 seconds when all DC parameters in the amp are stable. The latter two are high quality SDS relays, very expensive ones. I had them for ages in my drawer so this was the right opportunity to use them finally. Also, I added AC off detection circuit which should turn off the output relays when AC goes off (thus avoiding unwanted noise in speakers). As the final ingredient of the cake, I put a LED as indicator of protection state. This LED blinks when something goes wrong and it is very usefull for detecting what triggers the protection. (For example, when I turn on a vacuum cleaner, it often happens that protection circuit triggers as if AC is turned off).
Landing. Once the mounting board was finished it was the time to put it inside the box. Prior to "landing" i installed distancers of 1cm in height, which would provide space between the box bottom and the mounting board. I used eight of those for extra stability. When I finally landed the mounting board, I realised something terrible: heatsink ribs orientation was not suited for this box! Instead of going up-down, they go left-right.
It was a defeat of the moment indeed. After some thinking I got an idea to put a brushless cooler between the sinks in order to provide more air flow. I was afraid that this woud make noise and be heard in the speakers... This had to be tested...
Final touch. The box was designed to be very simple in design, with one power switch and a LED. On the sides of the box I decided to put wood. However, the carpenter did not understand me, instead of having 2cm on the outside of the box and 1cm of wood inside the box, he did it vice versa. This is ok, as I will cut the wood again, but this time by myself. I just need some free time...
Sound. Ah, this amp sounds great! 😀 Sansui ES-209 finally got the right driver. I am currently using Onkyo Integra p3060R preamp, which is a masterpeice indeed. For the source of the sound I use Sony CDP202es and tweaked Philips CD910. Both CDs work well with Studio. My fears were not in place, the cooler is not making any noise at all. When the sources of sound are not playing, it is amazing how silent the Studio is. If I put the potentiometer of preamp to high values, I can barely hear a little noise. Also, it seems that PIC is not interfering with the sound too. It is quiet in its place, maybe too fast with triggering the protection, but maybe it is better this way.
Fell free to ask what you want to know. Also, I am willing to provide schematics for the protection circuit, soft start and program for PIC (in mikroPascal). I just need some time to prepare it all.
Motivation. I have a pair of great loudspeakers by Sansui. The model is ES209, real rare beaty. They are powerfull speakers with 110W of music power. The main feature of this treasure is a passive radiator which is indeed capable of producing very low frequencies.
I have been using various amplifiers with these speakers and I was not satisfied. Simply, I felt there was more that they could give. I started with integrated amplifier SONY TA-FE370. It is one great amplifier with really good ratio value/quality. However, this amplifier did not fit well with Sansui speakers. The sound was too wide, too detailed and very sharp. I tried the amp with Yamaha speakers as well, but the situation was much better. The conclusion was either to change the speakers or to change the amp.
In the meantime, a friend provided me with two T-Amp amplifiers. I instantly made a powerfull stabilized power supply 2x 12V/4A. I connected T-Amps in biamp configuration and tested them with ES209. The effect was a very nice soft sound, very good quality indeed. Unfortunately, at large power T-Amps created distortion. Its nominal power is 10W, but one amp hardly produced 7W without distortion. Furthermore, the sensitivity of ES209 is around 88dB which required much stronger amplifier. T-Amp generally works well with speaker of >90 dB of senisitivity, so this idea failed as well.
After I explained all this trouble, Bora suggested me the Studio amplifier as an ultimate solution for Sansui ES209.
First steps. As I am DIY fanatic, I decided to make everything from scratch. The first task was to make PCBs and to test capabilities of the amplifier. It was quite easy to make PCBs, since it has no small footprints, such as IC or smd components. It is all discrete, true hole implementation. The next step was to get output transistors. I was lucky enough to find MJL21193 and MJL21194. I got them from a friend who has an electronics shop and he recomended the pair as very good for audio applications. As pre-output stage drivers I used MJE340 and MJE350.
Heatsink. Roaming over flee market I bumped on smuggler who was selling a full box of various heatsinks for a funny price of 20 euro. Of course, I did not miss the chance and bought them all. The box contained various sizes of unused heatsinks, brand new, aluminium. I selected two largest ones and mounted them on the transistors. Misfortune in the decision showed up later when I realized that the orientation of ribs is not fitting the box. I will explain this problem in details in few lines below.
Testing. After the first two boards were assembled I attached a regulater power supply +- 35V/1A. Few minutes were required to tune the quicient currents as design suggested. Then I attached PC CD player and SAL speakers, oscilloscope and a tone generator. Unfortunately, I realized (and heard) that the upper part of a positive semi period of a sinus signal has "hair", i.e. oscillations. On the speakers it sounded as if forest hums due to wind. I was very unhappy.
Finding the bug. Few days after agony and disapointment, I found a guy on a forum who suggested me to change all transistors. This was not making me happy, but I started the tedious work. Luckily for me, the first two transistors I changed were the input ones, BC546, situated in differential pair of the input stage. Instead of BC546 I put BC182 (the ones I had at hand) I turned the power back on and voila! All worked well! I left BC182, following the golden engineering rule "if it works- do not change it". I do not know what was causing the oscillations, but faulty transistors were probably the reason (later on I discarded 50 pcs of those- I did not want to see them again).
Box. I had the opportunity to use a workshop with various tools for metal handling. I asked the owner (who happened to be my father in law, btw.) for the possibility to make one nice metal box. The man was very happy, since he was very bored being in pension. The guy came up the very next day with a finished box outline, shaped on a aluminuim press. The thickness of the material he used was 1.125 mm. I was amazed how fast he made the skeleton. After this point I knew there was no return and that the amp had to be finished.
I decided to spend the whole week in the workshop, drilling, cutting and bending metal. I had never realized before that there was so much work in mechanics for one simple box.
The box was designed in such way that there are five fixed edges which are not moveable. They are welded to each other. The top cover is the only one designed to be able to open. This approach was telling me I would have trouble with mouting the elements inside the box. A logical decision was to make a metal plate which will hold all the electronics and that plate would be fixed for the bottom of the box after the amp is finished. It later showed to be a right decision.
Implementation. I got back to my working desk and started fixing heatsinks and other parts. I used silver wire 2.2mm^2 with teflon insulation. I spent four meters of the wire. I also used KCUG japanese capacitors on critical places. Also, in the filter section I put 4x10,000uF which are bypassed with 330nF/250V Siemens MKP capacitors. The transformer I used is custom made 2x30V AC, 300VA torus. Such transformer required soft start circuit.
For wiring AC power lines, I used a simple copper wire of 2.5mm^2. I also put a fuse f 1.6A (fast) on the primary of the transformer. Finally, the switch used is cutting both AC wires at the same time, thus resolving problems caused by large inductance of the primary.
Protection. I was concerned about my Sansui speakers, since the amplifier was supplied by +-43V DC and the transformer was capable of giving 5A per channel. This is a real danger for speakers. A logical conclusion was that a protection is a must to make.
Since I work with microcontrollers as a professional, I was immediately having an idea of using a PIC to control the soft start circuit and to make protection for speakers. However, I recall reading that PIC and other microcontrollers are not desirable in audio circuits especially on signal path. I was trying to respect this attitude and I designed a protection circuit which is not requiring a direct connection of the PIC to audio signal. The bacics of the protection are taken from this web page: http://sound.westhost.com/project33.htm
Variation of the circuit is made in such way that in collector of Q3 I put an opto-coupler which is then connected to PIC. In this way I separated PIC from audio signal path. I also decided to use a separate power supply for PIC and for the purpose I used a 220/12V transformer.
The protection circuit drives three relays. The first one is Schrak relay for the soft start circuit. It is turned off after 1 second. The other two relays are output relays for speakers. They turn on after 10 seconds when all DC parameters in the amp are stable. The latter two are high quality SDS relays, very expensive ones. I had them for ages in my drawer so this was the right opportunity to use them finally. Also, I added AC off detection circuit which should turn off the output relays when AC goes off (thus avoiding unwanted noise in speakers). As the final ingredient of the cake, I put a LED as indicator of protection state. This LED blinks when something goes wrong and it is very usefull for detecting what triggers the protection. (For example, when I turn on a vacuum cleaner, it often happens that protection circuit triggers as if AC is turned off).
Landing. Once the mounting board was finished it was the time to put it inside the box. Prior to "landing" i installed distancers of 1cm in height, which would provide space between the box bottom and the mounting board. I used eight of those for extra stability. When I finally landed the mounting board, I realised something terrible: heatsink ribs orientation was not suited for this box! Instead of going up-down, they go left-right.
It was a defeat of the moment indeed. After some thinking I got an idea to put a brushless cooler between the sinks in order to provide more air flow. I was afraid that this woud make noise and be heard in the speakers... This had to be tested...Final touch. The box was designed to be very simple in design, with one power switch and a LED. On the sides of the box I decided to put wood. However, the carpenter did not understand me, instead of having 2cm on the outside of the box and 1cm of wood inside the box, he did it vice versa. This is ok, as I will cut the wood again, but this time by myself. I just need some free time...
Sound. Ah, this amp sounds great! 😀 Sansui ES-209 finally got the right driver. I am currently using Onkyo Integra p3060R preamp, which is a masterpeice indeed. For the source of the sound I use Sony CDP202es and tweaked Philips CD910. Both CDs work well with Studio. My fears were not in place, the cooler is not making any noise at all. When the sources of sound are not playing, it is amazing how silent the Studio is. If I put the potentiometer of preamp to high values, I can barely hear a little noise. Also, it seems that PIC is not interfering with the sound too. It is quiet in its place, maybe too fast with triggering the protection, but maybe it is better this way.
Fell free to ask what you want to know. Also, I am willing to provide schematics for the protection circuit, soft start and program for PIC (in mikroPascal). I just need some time to prepare it all.
Consider if you could mount the heatsinks vertical - gives better cooling.
But nice work.
It looks finished, whereas my own (and many others') projects sometimes seem to have that little need-to-be-done-sometime-soon in it.
But nice work.
It looks finished, whereas my own (and many others') projects sometimes seem to have that little need-to-be-done-sometime-soon in it.
It looks very beautiful 🙂
One thing I thought of, did you ground the NC contacts of the speaker relays?
One thing I thought of, did you ground the NC contacts of the speaker relays?
Hi !
First of all, welcome on board !
Nice project but nothing spec.
Second, don't blame us, fins(heatsinks) have wrong angle,
Regards zeoN_Rider
First of all, welcome on board !
Nice project but nothing spec.
Second, don't blame us, fins(heatsinks) have wrong angle,
Regards zeoN_Rider
What they are saying is that those fins are horizontal while they should be vertical. In other words those heatsinks would be doing better job if those fins could be rotated in 90°.
However, go for it. Test it in real life, i.e. measure temperature on them and inside closed enclosure while amplifier is working serious job for 1/2-1 hour. If it satisfies you, than you're ok.
I have the same heatsinks, yet I have managed to obtain enough lengths to place them as enclosure sides, which means, that they are some 400 mm in length (each side) and placed in outside world. For normal listening levels (±1 building floor) it'll be just fine .😉
However, go for it. Test it in real life, i.e. measure temperature on them and inside closed enclosure while amplifier is working serious job for 1/2-1 hour. If it satisfies you, than you're ok.
I have the same heatsinks, yet I have managed to obtain enough lengths to place them as enclosure sides, which means, that they are some 400 mm in length (each side) and placed in outside world. For normal listening levels (±1 building floor) it'll be just fine .😉
- Status
- Not open for further replies.