The connections for Q15 and 16 you refer to are the heatsinks for those two devices. It shows them being grounded. Yes R38-R45, 73 and 74 are 274R. Q25 and 26 are 2N5550/2n5400 per the legend upper left.
Oops, didn't see Brlmat's post
I have been following the Thread since it started.
I referred to the latest sch in post331.
I can see an electrolytic (22uF) in the input that defines the LF passband.
Are you telling me that the schematic is wrong?
Last edited:
I don't get all this bandwidth discussion. I mean, yes, one usually uses a simple film capacitor in series with the input to limit the bandwidth, but doesn't using a preamp/ buffer do the same thing? They are both bandwidth limited. And almost any amp has a preamp or a buffer.
Hi Guys
Attached is an updated schematic and layout, to correct the BJT sub notes and the voltage rating for a few caps.
Andrew, the PA is given a wide bandwidth except for the rolloff caused by C11 and R56 in the feedback path. Because the original specification was made by Brlmat, where the amp is to be used for DJ purposes, hard use, continuous high output, used in conjunction with a limiter - I decided to place input band limiting on the No-clip sub board.
Not everyone building this amp needs the anti-clip circuit, so it is not a part of the basic circuit. Others may wish to modify the No-clip to simply be a distortion indicator.
Have fun
Kevin O'Connor
Attached is an updated schematic and layout, to correct the BJT sub notes and the voltage rating for a few caps.
Andrew, the PA is given a wide bandwidth except for the rolloff caused by C11 and R56 in the feedback path. Because the original specification was made by Brlmat, where the amp is to be used for DJ purposes, hard use, continuous high output, used in conjunction with a limiter - I decided to place input band limiting on the No-clip sub board.
Not everyone building this amp needs the anti-clip circuit, so it is not a part of the basic circuit. Others may wish to modify the No-clip to simply be a distortion indicator.
Have fun
Kevin O'Connor
Hi Guys
Just a note about resistor values:
I use 1% metal-films rated at 600mW. These were the Philips MRS25F-series from since the 80s or so. These became BC Components with a strange long series number.... but there are lots of manufacturers making equal quality parts. The foot print on the board is for the MRS25F.
Mouser has a range of MFs that are a bit larger. I believe even the leads are thicker. Thay may require elevation from the board to fit.
Most resistor positions in than amp can actually lower ratings than 600mW - 250mW or so. The exceptions that really should be 1/2W or 600mW are R16,26,35,36,50,54,55,57.
Values, like 274 can just as easily be 270. Both are available as 1%, but the 270 is a standard E24 number.
I'll be posting a full package of gerbers, excellon, description of circuit operation, testing, parts substitution, and other tidbits on my site today or tomorrow. Finally learned how to do the uploads there. The project name is different on my site, as "LTT4", which stands for "Lin Triple T 4 output-pairs". There will be many other project downloads to follow.
Have fun
Kevin O'Connor
Just a note about resistor values:
I use 1% metal-films rated at 600mW. These were the Philips MRS25F-series from since the 80s or so. These became BC Components with a strange long series number.... but there are lots of manufacturers making equal quality parts. The foot print on the board is for the MRS25F.
Mouser has a range of MFs that are a bit larger. I believe even the leads are thicker. Thay may require elevation from the board to fit.
Most resistor positions in than amp can actually lower ratings than 600mW - 250mW or so. The exceptions that really should be 1/2W or 600mW are R16,26,35,36,50,54,55,57.
Values, like 274 can just as easily be 270. Both are available as 1%, but the 270 is a standard E24 number.
I'll be posting a full package of gerbers, excellon, description of circuit operation, testing, parts substitution, and other tidbits on my site today or tomorrow. Finally learned how to do the uploads there. The project name is different on my site, as "LTT4", which stands for "Lin Triple T 4 output-pairs". There will be many other project downloads to follow.
Have fun
Kevin O'Connor
Hi Guys
Wayne, you did not look very deep into the site, did you? I also said today or tomorrow.
The bulk of what is on my site is aimed at MI and in that vein has an emphasis on tube circuitry. There are lots of accessories that are solid state, like the Power Scaling kits, power supplies, Quiet Supply active hum filters, and all the switching circuits. Our MI books are industry standards because they look beyond MI into hifi, robotics, and further, to see what can be applied to audio.
Some of the books are about general electronics and include hifi. We are working on a book series dedicated to hifi, called Audio Enlightenment, the first volumes of which should be out by year's end, with the first four volumes already mapped out and lots of content already in place. The first volume is an overview of audio systems, grounding, feedback, acoustics. The second volume is tube oriented, so that the first two volumes together can satisfy what was originally to be the 'Designer Notes" book. The third volume is solid-state power amps designs. The fourth is about head phone driver circuits, with a lot of info about ESL phones.
Because of how google searches and rankings work, and other issues of the online store portion of the site, much of what is obvious about the site is "tube" oriented. We are working on an html front page that will help to show more of the breadth of what we actually do, with the hifi part more obvious.
The Design Download section we are starting up is almost all hifi designs: power amps, cross overs, and other things, the bulk of which are solid-state.
Have fun
Kevin O'Connor
Wayne, you did not look very deep into the site, did you? I also said today or tomorrow.
The bulk of what is on my site is aimed at MI and in that vein has an emphasis on tube circuitry. There are lots of accessories that are solid state, like the Power Scaling kits, power supplies, Quiet Supply active hum filters, and all the switching circuits. Our MI books are industry standards because they look beyond MI into hifi, robotics, and further, to see what can be applied to audio.
Some of the books are about general electronics and include hifi. We are working on a book series dedicated to hifi, called Audio Enlightenment, the first volumes of which should be out by year's end, with the first four volumes already mapped out and lots of content already in place. The first volume is an overview of audio systems, grounding, feedback, acoustics. The second volume is tube oriented, so that the first two volumes together can satisfy what was originally to be the 'Designer Notes" book. The third volume is solid-state power amps designs. The fourth is about head phone driver circuits, with a lot of info about ESL phones.
Because of how google searches and rankings work, and other issues of the online store portion of the site, much of what is obvious about the site is "tube" oriented. We are working on an html front page that will help to show more of the breadth of what we actually do, with the hifi part more obvious.
The Design Download section we are starting up is almost all hifi designs: power amps, cross overs, and other things, the bulk of which are solid-state.
Have fun
Kevin O'Connor
Hi Guys
The output inductor is in uH, so not a real audio bandwidth shaper. Just for stability with loads.
Have fun
Kevin O'Connor
The output inductor is in uH, so not a real audio bandwidth shaper. Just for stability with loads.
Have fun
Kevin O'Connor
Kevin,
Wonderful work and web site as well
MRS25F, reminds me of buying parts from Electrosonic. Lots of replacements however,
Digi-Key still sells MRS25F
http://www.digikey.ca/product-detail/en/MRS25000C2740FRP00/PPC274ZCT-ND/594984
Another note, I use Mouser project manager for my BOMs. You can share them as well. Some designs have over 100+ different comps, so it can take a fair bit of time to input all this data. I structure my CAD DB to o/p vendor/supplier PN's, to minimize this backend work of BOM load/maintenance. Another good thing about having BOM's loaded, is that you get discontinuance notices.
Digi-Key has something similar. I cross source parts from both these suppliers. At anyone time a part can be out of stock and on back-order, this can lead to long delays at times.
God I think of the manual work that had to be done in the old days.
Cheers
Rick Savas
Wonderful work and web site as well
MRS25F, reminds me of buying parts from Electrosonic. Lots of replacements however,
Digi-Key still sells MRS25F
http://www.digikey.ca/product-detail/en/MRS25000C2740FRP00/PPC274ZCT-ND/594984
Another note, I use Mouser project manager for my BOMs. You can share them as well. Some designs have over 100+ different comps, so it can take a fair bit of time to input all this data. I structure my CAD DB to o/p vendor/supplier PN's, to minimize this backend work of BOM load/maintenance. Another good thing about having BOM's loaded, is that you get discontinuance notices.
Digi-Key has something similar. I cross source parts from both these suppliers. At anyone time a part can be out of stock and on back-order, this can lead to long delays at times.
God I think of the manual work that had to be done in the old days.
Cheers
Rick Savas
Last edited:
Sorry to be annoying but for this newbee, can you explain the audio difference between 1% and 5% resistors. I see a lot of build docs that say it doesn't matter and then others that stress 1%.
Thanks, Terry
Thanks, Terry
Most 1 percent are metal film, and most 5 percent are carbon film,
I guess some prefer metal film resistors,
Sometimes matching of transistors are important, I guess the corresponding resistors matched closer helps.
I guess some prefer metal film resistors,
Sometimes matching of transistors are important, I guess the corresponding resistors matched closer helps.
At one time 1% R's cost a premium, nowaday, there is very little in $ diff. I use 1% all the time regardless if I need that accuracy or not.
The ckt design dictates the tolerance of the R along with the other parameters..
Filters, or ckts that require tight balance use 1% or better.
I built a LCR meter of late and used 0.01% R's since they were part of the calibration adjustment of the instrument.
The ckt design dictates the tolerance of the R along with the other parameters..
Filters, or ckts that require tight balance use 1% or better.
I built a LCR meter of late and used 0.01% R's since they were part of the calibration adjustment of the instrument.
Hi Guys
As mentioned, the tolerance of a resistor value used to be indicative of the type of resistor, and before that of how tight the manufacturing process was. You could get carbon film 5% that when measured were more accurate to their value than ones that were nominally better by their markings.
I haven't used carbon resistors since 1990. Carbon resistors have several distortion mechanisms that make them a poor choice if you want low-THD. Bryston found that if the feedback R was carbon, its distortion swamped the entire contribution of the rest of the circuitry, so they went wholly to MF.
Metal-film is as close to a perfect resistor at audio frequencies and at acceptable cost as you can find. So, the odd values are a reflection of this in my schematics.
Even in a guitar amp, where carbon distortion is maximised and can contribute a "sweetness", it becomes too much when designing a circuit to distort - turning to mud - and has the other thermal issues that make unit-to-unit reliability and consistency impossible. So, I use MF there and design the circuit for the sweetness or tone I want.
Others will tell you that if you use carbon within its limit, it will be great. That is true with all parts. However, carbon has such a mix of vaguaries of behaviour that I just don't want them in my builds. You can build the amp here with carbons and may find you prefer it to one you build with MF. It's always your choice.
Have fun
Kevin O'Connor
As mentioned, the tolerance of a resistor value used to be indicative of the type of resistor, and before that of how tight the manufacturing process was. You could get carbon film 5% that when measured were more accurate to their value than ones that were nominally better by their markings.
I haven't used carbon resistors since 1990. Carbon resistors have several distortion mechanisms that make them a poor choice if you want low-THD. Bryston found that if the feedback R was carbon, its distortion swamped the entire contribution of the rest of the circuitry, so they went wholly to MF.
Metal-film is as close to a perfect resistor at audio frequencies and at acceptable cost as you can find. So, the odd values are a reflection of this in my schematics.
Even in a guitar amp, where carbon distortion is maximised and can contribute a "sweetness", it becomes too much when designing a circuit to distort - turning to mud - and has the other thermal issues that make unit-to-unit reliability and consistency impossible. So, I use MF there and design the circuit for the sweetness or tone I want.
Others will tell you that if you use carbon within its limit, it will be great. That is true with all parts. However, carbon has such a mix of vaguaries of behaviour that I just don't want them in my builds. You can build the amp here with carbons and may find you prefer it to one you build with MF. It's always your choice.
Have fun
Kevin O'Connor
Hi guys!
I've read that replacing the input bipolar with a film cap is better for the sound quality. Is this true?
Also how do you calculate the input filter's cut-off frequency? And so it's clear, the input filter is the cap in series with the hot wire and the resistor across the hot wire and ground, right?
Is the formula this one?
f = 1/(2*pi*R*C)
If so, is a 0.6hz cutoff frequency normal for a equalizer?
Do you always have to use this input cap? For example i'll have two buffers, one per channel, which will then mix into one channel which will go into a variable low pass filter. Is it ok if i use the input cap (4.7u film) only before the low pass filter with a 47k input impedance? Like having only one filter, between the buffers and the low pass.
And also, using this formula, would the cut off frequency for this amp be 0.3hz? Or 0.15? How do you calculate it if you have a resistor across, than a cap in series with a low value resistor and than one more resistor across?
I've read that replacing the input bipolar with a film cap is better for the sound quality. Is this true?
Also how do you calculate the input filter's cut-off frequency? And so it's clear, the input filter is the cap in series with the hot wire and the resistor across the hot wire and ground, right?
Is the formula this one?
f = 1/(2*pi*R*C)
If so, is a 0.6hz cutoff frequency normal for a equalizer?
Do you always have to use this input cap? For example i'll have two buffers, one per channel, which will then mix into one channel which will go into a variable low pass filter. Is it ok if i use the input cap (4.7u film) only before the low pass filter with a 47k input impedance? Like having only one filter, between the buffers and the low pass.
And also, using this formula, would the cut off frequency for this amp be 0.3hz? Or 0.15? How do you calculate it if you have a resistor across, than a cap in series with a low value resistor and than one more resistor across?
Last edited:
The input high pass filter is calculated using the R & C values in the equation:
F-3dB = 1 / Pi / 2 / R / C
Using the R value of 47k (E24 1% value) and replacing the 22uF electrolytic with a 2u2F film, you get F-3dB = 1/2/3.142/47000/0.0000022 ~=1.5Hz
This gives F-1dB ~3Hz
Using 22uF gains nothing in sound performance.
Now that the input high pass has been set, use the relationship:
Cnfb >= 1.5 * Cin * Rin / Rnfb >= 1.5 * 2u2F * 47k / 1k >= 155uF. Use 220uF
With the correct ratio of Cin:Cnfb, using 2m2F to 4m7F achieves nothing in sound quality performance. That Ratio requirement is what D.Self investigated when he showed that using too small a Cnfb value led to increased and avoidable THD.
For PA duty, the F-3dB frequency can be increased by at least two octaves and still get a very strong 30Hz to 40Hz performance that meets the expectations of the paying audience.
Further, most PA include a much steeper and higher frequency roll off (F-3dB) high pass filter to protect the speaker/s from abuse.
Expect a 3pole or higher set to somewhere around 20Hz to 40Hz as such protection. This protective filter completely invalidates the requirement for the 22uF & 2m2F capacitors specified in the schematic.
F-3dB = 1 / Pi / 2 / R / C
Using the R value of 47k (E24 1% value) and replacing the 22uF electrolytic with a 2u2F film, you get F-3dB = 1/2/3.142/47000/0.0000022 ~=1.5Hz
This gives F-1dB ~3Hz
Using 22uF gains nothing in sound performance.
Now that the input high pass has been set, use the relationship:
Cnfb >= 1.5 * Cin * Rin / Rnfb >= 1.5 * 2u2F * 47k / 1k >= 155uF. Use 220uF
With the correct ratio of Cin:Cnfb, using 2m2F to 4m7F achieves nothing in sound quality performance. That Ratio requirement is what D.Self investigated when he showed that using too small a Cnfb value led to increased and avoidable THD.
For PA duty, the F-3dB frequency can be increased by at least two octaves and still get a very strong 30Hz to 40Hz performance that meets the expectations of the paying audience.
Further, most PA include a much steeper and higher frequency roll off (F-3dB) high pass filter to protect the speaker/s from abuse.
Expect a 3pole or higher set to somewhere around 20Hz to 40Hz as such protection. This protective filter completely invalidates the requirement for the 22uF & 2m2F capacitors specified in the schematic.
I contribute here at my peril as I have not read all of the posts, and I can even less recall that many!
I would like to state that I am in agreement with Kevin not only as far as quality of his explanations go but also his astute work on this project. The following should in no way detract from that.
I hope that is accepted.
I can do nothing else though than agree with Andrew's last posts. I honestly need to be convinced what one wants in audio with a -1db point of lower than 10 Hz. Has l.f. hearing substantially improved over the last decade? Or loudspeakers? Or are we chasing specs? (OK, there are environments that can go that low - in a manner of speaking - I say environments because reproduction of such frequencies has to do as much with the environment as with the drivers.) Golly; I can wave my hand at a considerable amplitude at 10 Hz and hear nothing ....🙂
Granted that this might be a small point compared to the rest of the discussion. But then there was a lot about the influence on sound of electrolytics, and if one suddenly has to go 100x higher to obviate that when one runs out of the range of polyester/polypropylene, I would humbly ask for some explanation of the importance of the sub-10Hz region. It is not that there are not advantages of response staying out of there.
My apology if I had missed that somewhere earlier.
I would like to state that I am in agreement with Kevin not only as far as quality of his explanations go but also his astute work on this project. The following should in no way detract from that.
I hope that is accepted.
I can do nothing else though than agree with Andrew's last posts. I honestly need to be convinced what one wants in audio with a -1db point of lower than 10 Hz. Has l.f. hearing substantially improved over the last decade? Or loudspeakers? Or are we chasing specs? (OK, there are environments that can go that low - in a manner of speaking - I say environments because reproduction of such frequencies has to do as much with the environment as with the drivers.) Golly; I can wave my hand at a considerable amplitude at 10 Hz and hear nothing ....🙂
Granted that this might be a small point compared to the rest of the discussion. But then there was a lot about the influence on sound of electrolytics, and if one suddenly has to go 100x higher to obviate that when one runs out of the range of polyester/polypropylene, I would humbly ask for some explanation of the importance of the sub-10Hz region. It is not that there are not advantages of response staying out of there.
My apology if I had missed that somewhere earlier.
Cherry made some point on this subject many years ago in Wireless World.
The basic thrust was the roll-off from the many stages in the recording/reproducing chain were cumulative, and thus detrimental.
An easy fix is to convert from a single pole input filter to a two-pole filter with a Q=1.0 around 10hz or so. The values of the caps will be such that a smallish film cap will do the job. NAD does this on a couple of models. If not worried about DC from the preceeding stage, just eliminate the input cap. The Leach amplifier, Adcom GFA555, etc, are built this way.
I once took a large mixing desk with 10 coupling caps per channel, upgraded the opamps, and jumped out 7 of the caps per channel. The remaining 3 caps were changed from electrolytic to film types. The new opamps were of the FET type so I wasn't as concerned about tthe changing source impedances.
The basic thrust was the roll-off from the many stages in the recording/reproducing chain were cumulative, and thus detrimental.
An easy fix is to convert from a single pole input filter to a two-pole filter with a Q=1.0 around 10hz or so. The values of the caps will be such that a smallish film cap will do the job. NAD does this on a couple of models. If not worried about DC from the preceeding stage, just eliminate the input cap. The Leach amplifier, Adcom GFA555, etc, are built this way.
I once took a large mixing desk with 10 coupling caps per channel, upgraded the opamps, and jumped out 7 of the caps per channel. The remaining 3 caps were changed from electrolytic to film types. The new opamps were of the FET type so I wasn't as concerned about tthe changing source impedances.
Leach Double Barrel Amplifier
I am looking forward to built the Leach Double barrel amplifier. Though i have built the Low Tim version and pleased with the sound signature.
I am planning to use two separate transformer for better channel separation i.e one transformer for each channel. As Late Marshall Leach used 12 Ampere of current in his Double Barrel Amp. I am not sure how much Current exactly required. How much should be the Maximum current.
Thinking to built the power supply according to the below link for each channel
Solid State Power Amplifier Supply Part 2.
Also would glad if someone can advise, in order to implement the Power supply as per the above link .Do i need to use a transformer with each Secondary winding 12 Ampere or each Secondary winding 6 Ampere.
By bridging the secondary winding in series, the voltage will increase but not sure about the current.
Best Regards
Saurav
I am looking forward to built the Leach Double barrel amplifier. Though i have built the Low Tim version and pleased with the sound signature.
I am planning to use two separate transformer for better channel separation i.e one transformer for each channel. As Late Marshall Leach used 12 Ampere of current in his Double Barrel Amp. I am not sure how much Current exactly required. How much should be the Maximum current.
Thinking to built the power supply according to the below link for each channel
Solid State Power Amplifier Supply Part 2.
Also would glad if someone can advise, in order to implement the Power supply as per the above link .Do i need to use a transformer with each Secondary winding 12 Ampere or each Secondary winding 6 Ampere.
By bridging the secondary winding in series, the voltage will increase but not sure about the current.
Best Regards
Saurav
- Status
- Not open for further replies.