Hi Palesha.
The schematic looks very familiar. I know I've seen it some where.
It's correct that it has some kind of overload protection, BUT, as far as I can see it's only protecting the output trannies. There's no protection of TR4. In case the protective circuit is activated, only D6 and R18 is limiting the current through TR4, and I don't like that, eventhough TR2, 3 and 4 are a lot cheaper than the output trannies.
Another thing, while writing my brain out about safety, has the output transistors got protective diodes build in?
And just another thing. I would have liked a collector-resistor on TR3, as a current-limiter. Just one of my own old-fashion ideas.
All that said, I too like the simplicity of it, and Palesha, if You decide to build it, do some experiments with C3, C5, C6 and C7. I suspect them all to have a big influence on the sound quality, but I recon You've already figured that out your self.
The schematic looks very familiar. I know I've seen it some where.
It's correct that it has some kind of overload protection, BUT, as far as I can see it's only protecting the output trannies. There's no protection of TR4. In case the protective circuit is activated, only D6 and R18 is limiting the current through TR4, and I don't like that, eventhough TR2, 3 and 4 are a lot cheaper than the output trannies.
Another thing, while writing my brain out about safety, has the output transistors got protective diodes build in?
And just another thing. I would have liked a collector-resistor on TR3, as a current-limiter. Just one of my own old-fashion ideas.
All that said, I too like the simplicity of it, and Palesha, if You decide to build it, do some experiments with C3, C5, C6 and C7. I suspect them all to have a big influence on the sound quality, but I recon You've already figured that out your self.
Yes,
u must have seen it at http://blackwidowaudio.com.
The circuit has two 2n3055 at the output. I was thinking of modifying it. There are no values given for the components.
The page is still incomplete. Hence i thought to have discussion about it.
Mahendra Palesha
u must have seen it at http://blackwidowaudio.com.
The circuit has two 2n3055 at the output. I was thinking of modifying it. There are no values given for the components.
The page is still incomplete. Hence i thought to have discussion about it.
Mahendra Palesha
Palesha,
It's part of a basic circuit I've seen at 70's in an application note of RCA. In the original project there are diodes in paralel (reversed) with output transistors, a Zobel network and a 5uH or 10uH inductor in series with the load. You have to take care about the thermal stability, but I remember this circuit is very simple and works well.
Regards,
It's part of a basic circuit I've seen at 70's in an application note of RCA. In the original project there are diodes in paralel (reversed) with output transistors, a Zobel network and a 5uH or 10uH inductor in series with the load. You have to take care about the thermal stability, but I remember this circuit is very simple and works well.
Regards,
One reason it looks familiar is that it is the first
solid state circuit from the original RCA transistor
manual. Later it became known as the HK Citation 12.
solid state circuit from the original RCA transistor
manual. Later it became known as the HK Citation 12.
Yeup! Nelson is right!
It could look like the RCA "70 Watt quasi complementary symmetry" amp on page 184, Fig. 263.
Try if you you can get your hands on RCA applicationmanual "PTA-400" from 1983.
I think you all will like this book!
Sonny
It could look like the RCA "70 Watt quasi complementary symmetry" amp on page 184, Fig. 263.
Try if you you can get your hands on RCA applicationmanual "PTA-400" from 1983.
I think you all will like this book!
Sonny
60 Watt Amp - Components
Hi, Guy's
The amp you are all talking about, is from my site Blackwidowaudio.
If you would like the component value's you sould have sent me an email and i would have posted them. I have now added the component list to the project articles
Also earlier in the thread someone said about changing C3, C5/6/7. Changing C3 will help the sound this could be removed even, or made a larger value. You could also add 100nF decoupling caps to the power supply rails.
And yes the output transistors are the only ones protected by the protection circuit. And no as far as i know there is no relation to this amp and Citation or RCA, the output configuration maybe similar thou. The Quasi-Comp was seen around in 70's
this amp is very simple and all considering is a very good performer
anyway hope that helps, and happy new year
Dan
Hi, Guy's
The amp you are all talking about, is from my site Blackwidowaudio.
If you would like the component value's you sould have sent me an email and i would have posted them. I have now added the component list to the project articles
Also earlier in the thread someone said about changing C3, C5/6/7. Changing C3 will help the sound this could be removed even, or made a larger value. You could also add 100nF decoupling caps to the power supply rails.
And yes the output transistors are the only ones protected by the protection circuit. And no as far as i know there is no relation to this amp and Citation or RCA, the output configuration maybe similar thou. The Quasi-Comp was seen around in 70's
this amp is very simple and all considering is a very good performer
anyway hope that helps, and happy new year
Dan
To BlackWidowAudio:
I've just visited your site again, and also looked again at your 60 Watt amplifier again - thanks for adding the component list this time - but I've to mention this:
you say on this page: "There are a few modifications you can do to this amp, for starters you could remove the input capacitor, but be warned only do this if you are sure pre-amp or driving stage, is protected from DC at it's output."; please add also when removing the input capacitor you may NEVER short circuit the input to the ground, it may lead to blown speakers and/or a blown diff. amp. Sometimes, the input is connected to the ground to test the amp on distortion or on DC-components at the output.
Best regards,
hugobross.
I've just visited your site again, and also looked again at your 60 Watt amplifier again - thanks for adding the component list this time - but I've to mention this:
you say on this page: "There are a few modifications you can do to this amp, for starters you could remove the input capacitor, but be warned only do this if you are sure pre-amp or driving stage, is protected from DC at it's output."; please add also when removing the input capacitor you may NEVER short circuit the input to the ground, it may lead to blown speakers and/or a blown diff. amp. Sometimes, the input is connected to the ground to test the amp on distortion or on DC-components at the output.
Best regards,
hugobross.
A good start
Palesha,
Comments? It is a good place to start. The circuit is ye olde so it could benefit greatly from a few modifications:
1. Remove R3,C2,R2,D1. Add a constant current diode between TR2/TR3 emitters and +40V.
2. Replace TR2/TR3 with an SSM2220P match pair.
3. Put a 470pF cap in parallel with R4 and a 4k7 in series with C3 (for high frequency stability reasons).
4. Remove C6,R1,R11. Connect a constant current diode (5mA to 10mA, 1W) from the R12,R13 node to +40V. Put a 47uF cap between the R21,R13 node and the R22,D4 node.
5. 2N3055s are very old, slow workhorses. You would be better to use more modern output transistors with much higher Ft, like MJ15022 or similar.
6. Add a 10-ohm resistor in series with a 220nF capacitor across the output to ensure the circuit is loaded at high frequencies (improves stability).
You can pick and choose which to make, they are independent. C7 is a critical component for the stability of the circuit. It should have a value of at least 47pF, perhaps 100pF for initia switch on. Otherwise the circuit may oscillate or oscillate when you connect a speaker to it. For best reults use tantalum electrolytics.
Palesha,
Comments? It is a good place to start. The circuit is ye olde so it could benefit greatly from a few modifications:
1. Remove R3,C2,R2,D1. Add a constant current diode between TR2/TR3 emitters and +40V.
2. Replace TR2/TR3 with an SSM2220P match pair.
3. Put a 470pF cap in parallel with R4 and a 4k7 in series with C3 (for high frequency stability reasons).
4. Remove C6,R1,R11. Connect a constant current diode (5mA to 10mA, 1W) from the R12,R13 node to +40V. Put a 47uF cap between the R21,R13 node and the R22,D4 node.
5. 2N3055s are very old, slow workhorses. You would be better to use more modern output transistors with much higher Ft, like MJ15022 or similar.
6. Add a 10-ohm resistor in series with a 220nF capacitor across the output to ensure the circuit is loaded at high frequencies (improves stability).
You can pick and choose which to make, they are independent. C7 is a critical component for the stability of the circuit. It should have a value of at least 47pF, perhaps 100pF for initia switch on. Otherwise the circuit may oscillate or oscillate when you connect a speaker to it. For best reults use tantalum electrolytics.
DC coupling to a pre-amp is the same as connecting to ground. Do not attempt DC coupling (to ground or a preamp) without some sort of offset adjustment!!! Your speakers will hate you for it (if they survive).
If the preamp has a capacitor (AC) coupled output you are fine... better check first.
If the preamp has a capacitor (AC) coupled output you are fine... better check first.
hugobross said:
My first Thought when i saw this Circuit was as others hear have stated from the RCA Transistor manual. I never Built the 70 Watt Circuit in the Manual. However at that time I had a Heathkit AR-1500 Receiver and i compaired the Schematic in the Power amp section of the AR-1500 to the RCA Circuit and Those were Identical down to the Transistor types. This Receiver was the Best available in it's day But if Pushed hard would Blow Output transistors. Heath came out with a revision to adress this Problem in the AR-1500A. I think that considering the instability Problems common in this type of output stage and the fact that Measurable amounts if Crossover Distortin is also a problem I would go with a Modern Pure Complimentry Design.
RCA Circuit
It's a good circuit, and yes, as Nelson says, the basis of the HK Citation 12 (didn't Nelson design this, in fact??).
I do not agree with Traderblam's option to use constant current sources. These are just not necessary on the input stage, as the input variation is tiny ocmpared to the output, and in fact because of their HF performance, a CCS feeding the voltage amplifier causes difficulties with stability. You want the collector load of the VA to increase at high frequencies (viz, collector load impedance to drop at HF), say 100KHz and beyond, so that the Bode/Nyquist stability criteria can be met with less lag compensation across the base/collector of the VA. A bootstrap does this just fine; they really are a good topology.
Tbm chooses to limit HF input with a cap from T1's base to ground, straddling R4. Generally, such rolloff is audible and better achieved by judicious use of lag compensation, two stage gain rolloff, and bootstrap falling impedance at HF.
The overcurrent protection takes no account of the CE voltage operating across the output devices; viz SOAR considerations. Nevertheless, it's a commendable and routinely used approach which causes minor clips on even low level music. While it protects the output stage commendably and is widely used in Pro-Audio amps like Fender BXRs, it does slightly compress the sound and in AB tests most can hear it. If you protect the drivers with a 10R base stopper on the outputs, and choose a driver which can source/sink up to 4A (such as the MJE15030/31), then the drivers will survive amp encounters with a rusty nail, and T4 will never be sacrificed.
A Zobel of 10R/100nF from output to ground will further improve stability under load at HF, and a 33pF cap from collector of TR4 to the feedback node will enhance stability without much affecting sonics.
This is a very good basis for an excellent SS amp, but for best sonics the protection should be eschewed, and some refinements made to the output stage such as a fully complementary design using highly linear bipolars like the 2SC3281/2SA1302.
Palesha, build it. The HK Citation 12 was very highly regarded in its day, and while the absence of current sources is not politically correct, this is a goody. I would rate it the Hispano-Suiza of the older topologies. Its minimalism and elegance, at least to me, is very appealing.
Cheers,
Hugh R. Dean
www.printedelectronics.com
It's a good circuit, and yes, as Nelson says, the basis of the HK Citation 12 (didn't Nelson design this, in fact??).
I do not agree with Traderblam's option to use constant current sources. These are just not necessary on the input stage, as the input variation is tiny ocmpared to the output, and in fact because of their HF performance, a CCS feeding the voltage amplifier causes difficulties with stability. You want the collector load of the VA to increase at high frequencies (viz, collector load impedance to drop at HF), say 100KHz and beyond, so that the Bode/Nyquist stability criteria can be met with less lag compensation across the base/collector of the VA. A bootstrap does this just fine; they really are a good topology.
Tbm chooses to limit HF input with a cap from T1's base to ground, straddling R4. Generally, such rolloff is audible and better achieved by judicious use of lag compensation, two stage gain rolloff, and bootstrap falling impedance at HF.
The overcurrent protection takes no account of the CE voltage operating across the output devices; viz SOAR considerations. Nevertheless, it's a commendable and routinely used approach which causes minor clips on even low level music. While it protects the output stage commendably and is widely used in Pro-Audio amps like Fender BXRs, it does slightly compress the sound and in AB tests most can hear it. If you protect the drivers with a 10R base stopper on the outputs, and choose a driver which can source/sink up to 4A (such as the MJE15030/31), then the drivers will survive amp encounters with a rusty nail, and T4 will never be sacrificed.
A Zobel of 10R/100nF from output to ground will further improve stability under load at HF, and a 33pF cap from collector of TR4 to the feedback node will enhance stability without much affecting sonics.
This is a very good basis for an excellent SS amp, but for best sonics the protection should be eschewed, and some refinements made to the output stage such as a fully complementary design using highly linear bipolars like the 2SC3281/2SA1302.
Palesha, build it. The HK Citation 12 was very highly regarded in its day, and while the absence of current sources is not politically correct, this is a goody. I would rate it the Hispano-Suiza of the older topologies. Its minimalism and elegance, at least to me, is very appealing.
Cheers,
Hugh R. Dean
www.printedelectronics.com
Hugh,
Nelson Pass published a MOSFET update to the HK Citation 12 in the 1980-81 time frame. While he expressed admiration for the original circuit at the time, he didn't claim credit for its design.
BTW, a significant difference between the RCA app-note circuit and the HK-12 is that HK chose to omit the active output stage protection, just as you have recommended.
Nelson Pass published a MOSFET update to the HK Citation 12 in the 1980-81 time frame. While he expressed admiration for the original circuit at the time, he didn't claim credit for its design.
BTW, a significant difference between the RCA app-note circuit and the HK-12 is that HK chose to omit the active output stage protection, just as you have recommended.
AKSA, it seems you and I are poles apart (pardon the pun 😛 ) on many points. This may be a reflection of the types of application that each of us designs for. My experience is the design of hifi amps where sound quality is the primary goal. In this context I beg to differ on some points you make:
"...constant current sources. These are just not necessary on the input stage, as the input variation is tiny ocmpared to the output, and in fact because of their HF performance, a CCS feeding the voltage amplifier causes difficulties with stability. "
In my experience the LTP is critical to performacne as it is where input and output are subtracted to form a error signal that is fed to the gain stage. The accuracy of this stage is really important - so small things matter a great deal here. I'm not sure what CCSs you have used that cause instability but I have used CCSs many, many times without issue.
"...A bootstrap does this just fine; they really are a good topology. "
The bootstrap in this circuit was surely just an old-fashioned way to mimmick a current source, except it doesn't work as well because the capacitor makes it frequency-dependent and the need for the Vbe across the darlingtons to vary non-linearly with the output current means the current source is not linear. From a small-signal point of view it is pretty much like running R11 from the collector of TR4 to the output - this isn't really what you want because TR4 (and the feedback loop) is effectively having to provide extra current to drive R11 in parallel with the darlingtons.
"...Tbm chooses to limit HF input with a cap from T1's base to ground, straddling R4. Generally, such rolloff is audible and better achieved by judicious use of lag compensation, two stage gain rolloff, and bootstrap falling impedance at HF. "
Not at all. The purpose is to prevent unwanted HF entering. This can come from the source or from noise associated with output current and can often cause instability. Perhaps this is why you have seen CCS in LTPs appearing to reduce stability? Keeping unwanted HF out of amp is very important in any amp design and it is not sensible to mimmick this using the closed loop response because the noise is still entering the system. A cap of this value will have negligible affect on the audio frequency response. However, allowing HF noise into the amp circuit will certianly degrade the sound quality.
"a 33pF cap from collector of TR4 to the feedback node will enhance stability without much affecting sonics. "
It may help with stability but it compromises the error correction of the feedback. As you know the feedback works by the amp measuring the difference between input and output and nothing else. By adding this cap a new signal is introduced that is not linearly-related to the output signal. The feedback will no longer minimize input-output error but some hybrid of input-output-TR4 collector voltage.
"for best sonics the protection should be eschewed, and some refinements made to the output stage such as a fully complementary design using highly linear bipolars like the 2SC3281/2SA1302. "
I'm pretty sure RCA chose two npn output devices because of the lack of availability of pnp complements at the time. It is still very difficult to find npn/pnp pairs that are well-matched. It can be done and certainly Krell do this but many contemporary circuits, like Naim's, choose a pseudo-complementary arrangement because it is much easier to match TR8/9 if they are the same device type and because faster npn devices are available than pnps.
The protection circuit is really there to protect your speakers. Insure your speakers before removing the protection circuit! 😉
"...constant current sources. These are just not necessary on the input stage, as the input variation is tiny ocmpared to the output, and in fact because of their HF performance, a CCS feeding the voltage amplifier causes difficulties with stability. "
In my experience the LTP is critical to performacne as it is where input and output are subtracted to form a error signal that is fed to the gain stage. The accuracy of this stage is really important - so small things matter a great deal here. I'm not sure what CCSs you have used that cause instability but I have used CCSs many, many times without issue.
"...A bootstrap does this just fine; they really are a good topology. "
The bootstrap in this circuit was surely just an old-fashioned way to mimmick a current source, except it doesn't work as well because the capacitor makes it frequency-dependent and the need for the Vbe across the darlingtons to vary non-linearly with the output current means the current source is not linear. From a small-signal point of view it is pretty much like running R11 from the collector of TR4 to the output - this isn't really what you want because TR4 (and the feedback loop) is effectively having to provide extra current to drive R11 in parallel with the darlingtons.
"...Tbm chooses to limit HF input with a cap from T1's base to ground, straddling R4. Generally, such rolloff is audible and better achieved by judicious use of lag compensation, two stage gain rolloff, and bootstrap falling impedance at HF. "
Not at all. The purpose is to prevent unwanted HF entering. This can come from the source or from noise associated with output current and can often cause instability. Perhaps this is why you have seen CCS in LTPs appearing to reduce stability? Keeping unwanted HF out of amp is very important in any amp design and it is not sensible to mimmick this using the closed loop response because the noise is still entering the system. A cap of this value will have negligible affect on the audio frequency response. However, allowing HF noise into the amp circuit will certianly degrade the sound quality.
"a 33pF cap from collector of TR4 to the feedback node will enhance stability without much affecting sonics. "
It may help with stability but it compromises the error correction of the feedback. As you know the feedback works by the amp measuring the difference between input and output and nothing else. By adding this cap a new signal is introduced that is not linearly-related to the output signal. The feedback will no longer minimize input-output error but some hybrid of input-output-TR4 collector voltage.
"for best sonics the protection should be eschewed, and some refinements made to the output stage such as a fully complementary design using highly linear bipolars like the 2SC3281/2SA1302. "
I'm pretty sure RCA chose two npn output devices because of the lack of availability of pnp complements at the time. It is still very difficult to find npn/pnp pairs that are well-matched. It can be done and certainly Krell do this but many contemporary circuits, like Naim's, choose a pseudo-complementary arrangement because it is much easier to match TR8/9 if they are the same device type and because faster npn devices are available than pnps.
The protection circuit is really there to protect your speakers. Insure your speakers before removing the protection circuit! 😉
Okay i have some words to.
Constant currentsources do raise CMRR in (I hate it!, I have typed "in the" as "int" for "integer" a lot of times) the diff. input stage. Wich again raise PSRR. Am i right or ????
The Lowpass filter on the input of a noninverting amplifier ( You can do it a inverting amplifier) prevent the amplifier from generating harmonics Even DC offset on the output. Try Capacitively couple a burst or a GSM style signal to the input of such a stage... You will be surprised!
Or your input circuit could work as a MIXER! if you don't use the lowpass filter.
Yes the CAP to ground can prevent a oscillation but then your input impedance maybe is to high! or the "Cobo" is multiplyed that much so you have to use the CAP to ground to lower the input impedance at high frequency.
The "bootstrap" thing adds another Frequency poll to the circuit.
B&O did have problems with this long time ago (So the story says) where they changed the CAP from MKT type to ELCAP type and the circuit stops oscillating.
It is okay to add a CAP as AKSA descriped. It help your circuit to get stable when driving Cap. or IND. load but i would do this just after the first VAS buffer not to load the VAS. If you have a high CAP or IND load at your outputstage will lower or raise your phasemargin. By adding this feedback you raise your phasemargin.
Zopel-network it is a mather of poor design/PCBlayout (Now i die!!!!). Your circuit should not get unstable at high frequency. Maybe your Phasemargin is to low?
I would go for outputdevices with a "flat" currentgain and a high Ft like the MJL1302/3281 or 2SA1302/SC3281 or MJL21193/94 or bla bla bla....
And it can be quassi stage. It can be really good to!!!
I think AKSA or BLACKWIDOW or PASS or Anthony's or RCA AMPS (I hope you all execute me if i have forgotten some!) are good designs. They sound different! YES they do ... What is your taste?
😉
Sonny
Constant currentsources do raise CMRR in (I hate it!, I have typed "in the" as "int" for "integer" a lot of times) the diff. input stage. Wich again raise PSRR. Am i right or ????
The Lowpass filter on the input of a noninverting amplifier ( You can do it a inverting amplifier) prevent the amplifier from generating harmonics Even DC offset on the output. Try Capacitively couple a burst or a GSM style signal to the input of such a stage... You will be surprised!
Or your input circuit could work as a MIXER! if you don't use the lowpass filter.
Yes the CAP to ground can prevent a oscillation but then your input impedance maybe is to high! or the "Cobo" is multiplyed that much so you have to use the CAP to ground to lower the input impedance at high frequency.
The "bootstrap" thing adds another Frequency poll to the circuit.
B&O did have problems with this long time ago (So the story says) where they changed the CAP from MKT type to ELCAP type and the circuit stops oscillating.
It is okay to add a CAP as AKSA descriped. It help your circuit to get stable when driving Cap. or IND. load but i would do this just after the first VAS buffer not to load the VAS. If you have a high CAP or IND load at your outputstage will lower or raise your phasemargin. By adding this feedback you raise your phasemargin.
Zopel-network it is a mather of poor design/PCBlayout (Now i die!!!!). Your circuit should not get unstable at high frequency. Maybe your Phasemargin is to low?
I would go for outputdevices with a "flat" currentgain and a high Ft like the MJL1302/3281 or 2SA1302/SC3281 or MJL21193/94 or bla bla bla....
And it can be quassi stage. It can be really good to!!!
I think AKSA or BLACKWIDOW or PASS or Anthony's or RCA AMPS (I hope you all execute me if i have forgotten some!) are good designs. They sound different! YES they do ... What is your taste?
😉
Sonny
"The protection circuit is really there to protect your speakers. Insure your speakers before removing the protection circuit!
"
Relays are speaker protection... the VI limiter keeps the output devices within their SOA curves so they don't die (hopefully!). If an output device dies closed circuit, the VI limiter is helpless and the speakers will die very quickly (which is where relays kick in). I guess they indirectly protect the speakers by protecting the OPS.. the extra clipping a crappy VI limiter may impose could damage the tweeters much easier than an amp with no VI limiting however.
"
Relays are speaker protection... the VI limiter keeps the output devices within their SOA curves so they don't die (hopefully!). If an output device dies closed circuit, the VI limiter is helpless and the speakers will die very quickly (which is where relays kick in). I guess they indirectly protect the speakers by protecting the OPS.. the extra clipping a crappy VI limiter may impose could damage the tweeters much easier than an amp with no VI limiting however.
RCA Power AMp
This is indeed the amplifier from the RCA transistor manual except the current limit circuit has been changed from the load line limiting that RCA had in the manual. I used to sell PCBs for this amplifier in the 70s and built dozens of them both with RCAs protetion cicuit (which was too agressive at times) and with this exact circuit.
This circuit was the basis of many solid state power amplifiers from the late 60s to late 70s when complementary designs took over.
This circuit tended to oscillate if driver transistors other than 40409s and 40410s were used. With other drivers, the open loop frequency response had to be reduced. However, this resulted in high levels of transient intermodulation distortion. This was the amplifier design that typified the "harsh transistor sound" that tube fans would put down and has beemn rendered obsolete by the much better sounding amplifier designs by Leach and others.
For some amp schematics, see www.schematicsforfree.com
This is indeed the amplifier from the RCA transistor manual except the current limit circuit has been changed from the load line limiting that RCA had in the manual. I used to sell PCBs for this amplifier in the 70s and built dozens of them both with RCAs protetion cicuit (which was too agressive at times) and with this exact circuit.
This circuit was the basis of many solid state power amplifiers from the late 60s to late 70s when complementary designs took over.
This circuit tended to oscillate if driver transistors other than 40409s and 40410s were used. With other drivers, the open loop frequency response had to be reduced. However, this resulted in high levels of transient intermodulation distortion. This was the amplifier design that typified the "harsh transistor sound" that tube fans would put down and has beemn rendered obsolete by the much better sounding amplifier designs by Leach and others.
For some amp schematics, see www.schematicsforfree.com
This amplifier uses only NPN output power transistors.
Both from positive and negative half of output.
You only need to buy 4 of them 2N3055 to build a stereo.
A completed amplifier and details can be found
http://www.blackwidowaudio.com/
Both from positive and negative half of output.
You only need to buy 4 of them 2N3055 to build a stereo.
A completed amplifier and details can be found
http://www.blackwidowaudio.com/
- Status
- Not open for further replies.