Never noticed unacceptable thumps....
On the other hand: there is a cut-off frequency due to C3 and the impedance of the boots-trap circuit, limiting the effectiveness of the circuit at lower frequencies and increasing distortion.
So maybe they got it correct
You also need to look at the input transistor at switch on and make sure there is no tendency for a reverse bias across the B-E junction to occur (which can degrade the noise figure of the transistor if it happens).
As a matter of interest, the Quad 303 ? (the old AC coupled one) uses a 47uf cap but I think the impedances are a little lower. So same design philosophy.
With a possible exception of the voltage amplifier transistor, I don't think that substituting modern parts will greatly improve the sound of this old design. I built a similar magazine project based on a Philips/Mullard design much like the 40W version, as long ago as 1971. This version used a BFY50 as VA transistor but it never reached great sound quality - until more recently when I fitted a spare Hitachi 2SD669 transistor. Then it sang like a bird, or so I thought .
I've since learned here that original BD139/40 were lower capacitance (Cob) types compared to later production and also the many copies from the likes of Onsemi, Fairchild, STmicro, Unisonic, CDIL and several Chinese sources. Some parts are now just general purpose power transistors for relays, regulators etc. and can sound dead here in this application.
The later Philips parts (perhaps after 1980?) were no longer the same specification referred to for the Philips 15-25-40W designs. The most critical to sound and stability is the voltage amplifier and its 27pF compensation cap. It would make sense to reconsider the original 27pF value if using "slower" copies or "faster" substitute types there. However, this calls for an oscilloscope since we are aware (I hope) that you can quickly destroy an amplifier by not having enough frequency compensation capacitance for stability.
Some folk become adept with making and using simple oscillation detectors as shown occasionally here by DiyAudio members. Worth a search and trying? Just make sure your amplifier is always operating stably first.
.I've since learned here that original BD139/40 were lower capacitance (Cob) types compared to later production and also the many copies from the likes of Onsemi, Fairchild, STmicro, Unisonic, CDIL and several Chinese sources. Some parts are now just general purpose power transistors for relays, regulators etc. and can sound dead here in this application.
The later Philips parts (perhaps after 1980?) were no longer the same specification referred to for the Philips 15-25-40W designs. The most critical to sound and stability is the voltage amplifier and its 27pF compensation cap. It would make sense to reconsider the original 27pF value if using "slower" copies or "faster" substitute types there. However, this calls for an oscilloscope since we are aware (I hope) that you can quickly destroy an amplifier by not having enough frequency compensation capacitance for stability.
Some folk become adept with making and using simple oscillation detectors as shown occasionally here by DiyAudio members. Worth a search and trying? Just make sure your amplifier is always operating stably first.
Dear Ian,
Thank you for the insight and observation reg the transistors, even other devices. perfectly working, established designs, also refuse to work, possibly due to this reason. Remember Bob Vidlar at NS? several times we find this "not up to the mark" component spoiling all the fun in a good ckt. Good to know you through this. I work mostly in the power supplies, new designs production as well as repairs of the power supplies whose parents are dead! Audio is my hobby. But I shall restart my forgotten experiments with small RF circuits, with your observations in mind. Keep in touch.
nmosfet
Thank you for the insight and observation reg the transistors, even other devices. perfectly working, established designs, also refuse to work, possibly due to this reason. Remember Bob Vidlar at NS? several times we find this "not up to the mark" component spoiling all the fun in a good ckt. Good to know you through this. I work mostly in the power supplies, new designs production as well as repairs of the power supplies whose parents are dead! Audio is my hobby. But I shall restart my forgotten experiments with small RF circuits, with your observations in mind. Keep in touch.
nmosfet
HI
I still recommend BD139/140. I've built several amps with these, mostly ST and ON Semi. I was suspicious of some types early on because of the similar TO-126 packaged devices using epitaxial base technology with ft's in the few MHz region. Before using any I checked the gain at 1MHz and if they were still around 100 I would use them. I did not find any low fT devices, but that does not mean to say that there might not be fakes out there.
And I have also built several amplifiers which do not use Miller capacitors (27pF or otherwise) which simulate and measure as having lower distortion than the usual standard circuits.
Most of the old style designs can be configured to use the epitaxial 2N3055 and MJ2955. However, even ST have announced retirement but TIP3055/MJ2955 might be a suitable replacement. As of writing I believe ON Semi still offer the metal cans.
I still recommend BD139/140. I've built several amps with these, mostly ST and ON Semi. I was suspicious of some types early on because of the similar TO-126 packaged devices using epitaxial base technology with ft's in the few MHz region. Before using any I checked the gain at 1MHz and if they were still around 100 I would use them. I did not find any low fT devices, but that does not mean to say that there might not be fakes out there.
And I have also built several amplifiers which do not use Miller capacitors (27pF or otherwise) which simulate and measure as having lower distortion than the usual standard circuits.
Most of the old style designs can be configured to use the epitaxial 2N3055 and MJ2955. However, even ST have announced retirement but TIP3055/MJ2955 might be a suitable replacement. As of writing I believe ON Semi still offer the metal cans.
Dear John,
Thank you for your observations. I have also tried with TIP 3055 and MJ 2955. Those sing very well, and I looked for a better device because I knew that the standard 2n3055 was a low FT device. After using the above, i changed the tone control to LF 356 circuit as given in an NS app note. Wonderful results. I have stopped using metal cans 20 years back, when the TOP-3 and TO-247 packages became available.
Recently I am having some device problems with TDA 1524 TC . despite a lot of checking, cant find the fault. probably the IC itself. Fakes or rejects must be finding their way in to market. Thanks again, and keep posting.
regards
nmosfet
Thank you for your observations. I have also tried with TIP 3055 and MJ 2955. Those sing very well, and I looked for a better device because I knew that the standard 2n3055 was a low FT device. After using the above, i changed the tone control to LF 356 circuit as given in an NS app note. Wonderful results. I have stopped using metal cans 20 years back, when the TOP-3 and TO-247 packages became available.
Recently I am having some device problems with TDA 1524 TC . despite a lot of checking, cant find the fault. probably the IC itself. Fakes or rejects must be finding their way in to market. Thanks again, and keep posting.
regards
nmosfet
Hi´Pabloso
I have a paper version of that book... Want a copy???
Also a lot of experience in this special version.
brand new PCB layout, relay controlled preamp etc.etc.
Have build more than 100 pcs of it..
Best regards
Tamas Kiralyfalvy
Hi´
Your "HISS" problem is very known in this design, Check your mid-voltage, it should be EXACT half of your supply voltage (you can add a trimmer in series with the 150kohm + 47kohm resistor that goes from + rail to the first transistor basis to adjust the mid voltage.
in the original design there is also a 470 ohm series resistor at the basis of the driver transistors, the series resistor at npn transistor basis should be paralled with a ex. 220nF poly cap, to stop oscillation.
In most cases this will stop the hiss. (it did for me, as I found oscillation here with the old PCB design).
I have created a lot of mod´s for this amplifier, also new PCB designs, for both preamplifier with relay controlled inputs (avoid a lot of cables).
Better noise performance, and easier control.
Tamas
Hi Reji
Your "hiss" problem is (from my recollection) due to the high gain of the amplifier and relatively high (almost open circuit) input impedance. The NFB is set by a 1.8k plus 4.7k in parallel =1.3k and the 33 ohm feedback ground resistor = about 40x. Most amplifiers have a gain of around 20-30x, which will help a little.
The input bias resistors of 150k give an equivalent 75k which is quite noisy. You should notice less hiss when connected to a low impedance preamp. Preferably the input bias stage should be built around a simple resistor divider using e.g. 27k and 22k (to pull the base voltage down slightly so the centre rail is centred) - but you may need 24k and 22k ... or a trimmer). Decouple the tap to ground (another advantage is that this arrangement only needs a 40V capacitor instead of 63V or 100V) then bias the base using a 10k (up to 22k) resistor. This should reduce the noise. THe optimum source resistance for the PNP input transistor is around 5k, but this needs to be fed from the preamp.
Your "hiss" problem is (from my recollection) due to the high gain of the amplifier and relatively high (almost open circuit) input impedance. The NFB is set by a 1.8k plus 4.7k in parallel =1.3k and the 33 ohm feedback ground resistor = about 40x. Most amplifiers have a gain of around 20-30x, which will help a little.
The input bias resistors of 150k give an equivalent 75k which is quite noisy. You should notice less hiss when connected to a low impedance preamp. Preferably the input bias stage should be built around a simple resistor divider using e.g. 27k and 22k (to pull the base voltage down slightly so the centre rail is centred) - but you may need 24k and 22k ... or a trimmer). Decouple the tap to ground (another advantage is that this arrangement only needs a 40V capacitor instead of 63V or 100V) then bias the base using a 10k (up to 22k) resistor. This should reduce the noise. THe optimum source resistance for the PNP input transistor is around 5k, but this needs to be fed from the preamp.
The book: "Audio amplifier systems"
Howdy everybody.
I have the book Audio amplifier systems. I have scanned it to 11 pdf files of about 10 Mb each. The book is very hard to find anymore.
The scanned resolution is 300 pixels. I took care to ensure the pcb pictures were flat on the glass. One should be able to use the pager to fabricate pcb.s.
If anyone wants the pdf.s contact me at:
eugenepohjola@msn.com
Best regards
Howdy everybody.
I have the book Audio amplifier systems. I have scanned it to 11 pdf files of about 10 Mb each. The book is very hard to find anymore.
The scanned resolution is 300 pixels. I took care to ensure the pcb pictures were flat on the glass. One should be able to use the pager to fabricate pcb.s.
If anyone wants the pdf.s contact me at:
eugenepohjola@msn.com
Best regards
The 40 Watt PA
Howdy again.
I remeber having built the 40 Watt version in the mid seventies and having had some problems with the noise.
The PA was quiet used with the Universal Preamplifier but when I inserted RC hi and lo -pass circuits in between the noise reappeared. Clearly the PA input needs to be connected to a low output impedance preamplifier. I seem to remember that simply connecting an external loading resistor of 10k between the input and minus quenched the noise.
My skills are not good enough to fiddle with the resistors of the internal circuitry to kill the noise but should I build the PA today I would do these modifications.
- I would always use plastic capacitors in signal propagation.
- I would enlarge the output el. cap. to 10 000 uF.
- I would use true complementary output configuration 2N3055(NPN), MJ2955(PNP).
In my blog eugpoh23.blogspot.com I have a study of a 50 Watt PA of my own design. Any feedback is warmly appreciated.
Best regards.
Howdy again.
I remeber having built the 40 Watt version in the mid seventies and having had some problems with the noise.
The PA was quiet used with the Universal Preamplifier but when I inserted RC hi and lo -pass circuits in between the noise reappeared. Clearly the PA input needs to be connected to a low output impedance preamplifier. I seem to remember that simply connecting an external loading resistor of 10k between the input and minus quenched the noise.
My skills are not good enough to fiddle with the resistors of the internal circuitry to kill the noise but should I build the PA today I would do these modifications.
- I would always use plastic capacitors in signal propagation.
- I would enlarge the output el. cap. to 10 000 uF.
- I would use true complementary output configuration 2N3055(NPN), MJ2955(PNP).
In my blog eugpoh23.blogspot.com I have a study of a 50 Watt PA of my own design. Any feedback is warmly appreciated.
Best regards.
Philips Circuits
Sorry, just arrived on the scene very recently.
From the discussions on this forum, it looks like many ( including me) are satisfied with the superb performance of Philips pre and power amps whilst some are not. The Philips book takes us thru a lot of technicalities.
However, it seems that some have faced noise problems etc .There are at least two major sources of noise that are generated due to earthing problems. For that the precaution that is to be taken is that while connecting preamp output to the Power amp the shield must be earthed at the earth end of the power amp and not preamp. Secondly the correct earthing point of power amp is the emitter point of the BD182.Full details of stray current loops and earthing are given on pages 71,72,73 of the book. Earthing of Universal preamp is also shown very precisely on page 42. Hopefully this should resolve the problems of noise.
There seems to be another issue of high value coupling capacitor with the speaker. There is a fear that this will lead to losses in low frequency .While this is true, it will be seen that the bass gain of Universal preamp is 20 db which is quite high and compensates for this loss .This is why the system did not seem to loose on bass. Other preamps used to just have around 10 to 12 db bass gain. I had used Philips components including the drivers and was very satisfied with the result. My recommendation is that please construct as per instructions in the book and use only Philips parts including the speaker drivers with 3 way sealed enclosures not forgetting their cross overs.
Another way to look at it is that there is a certification by Philips of meeting with the European standards for almost all circuits .We can rely on their integrity and I am sure they would have tested all the circuits before coming out with the book which is on global circulation and Bible for most of us. So if anything to the contrary we have to recheck our construction.
Chaks
Sorry, just arrived on the scene very recently.
From the discussions on this forum, it looks like many ( including me) are satisfied with the superb performance of Philips pre and power amps whilst some are not. The Philips book takes us thru a lot of technicalities.
However, it seems that some have faced noise problems etc .There are at least two major sources of noise that are generated due to earthing problems. For that the precaution that is to be taken is that while connecting preamp output to the Power amp the shield must be earthed at the earth end of the power amp and not preamp. Secondly the correct earthing point of power amp is the emitter point of the BD182.Full details of stray current loops and earthing are given on pages 71,72,73 of the book. Earthing of Universal preamp is also shown very precisely on page 42. Hopefully this should resolve the problems of noise.
There seems to be another issue of high value coupling capacitor with the speaker. There is a fear that this will lead to losses in low frequency .While this is true, it will be seen that the bass gain of Universal preamp is 20 db which is quite high and compensates for this loss .This is why the system did not seem to loose on bass. Other preamps used to just have around 10 to 12 db bass gain. I had used Philips components including the drivers and was very satisfied with the result. My recommendation is that please construct as per instructions in the book and use only Philips parts including the speaker drivers with 3 way sealed enclosures not forgetting their cross overs.
Another way to look at it is that there is a certification by Philips of meeting with the European standards for almost all circuits .We can rely on their integrity and I am sure they would have tested all the circuits before coming out with the book which is on global circulation and Bible for most of us. So if anything to the contrary we have to recheck our construction.
Chaks
A higher value coupling cap would extend LF response compared to one of lower value, not curtail it.