Hi Andy. Winter is around the corner and I thought maybe I will build the KX2 to keep me busy & warm!
Looking at he BOM, some parts I have others I need to buy. I bought a large heatsink on Ebay 370x50x280 for £40. Chopped it in half using a Tile Saw. Looks Big
. That bits sorted !
For a start I want to build the Amplifiers up and try them using parts I have at hand for the PSU's. Then I would spend money on nice cases etc..
Wonder if you could advise and answer a few questions for me please, regarding your deign.
I intend to build mono-blocks. I have four 10000uf caps. Would 10000uf per rail be enough? As the current draw would be half that of a single PSU?
I do have a pair of 200VA 19-9-19 Toloroids which may get me away to try the amp (Not screened so not ideal). But again lower magnetic fields as half the current. I did read where larger Toloroids are preferred. Apparently better for reducing the transfer of noise and RFI. So maybe swop out later..
I see you use Bridges rather than high current Hyper-Fast switching Diodes for the PSU. I have some of these so maybe I will use those instead and but snubber caps across them. Again some say bridges are better though?
I guess the 20A bridge earth lifter needs to be 20Amp to withstand mains shorts to earth?
The resistors you specified are expensive as you need to buy 100 off. I have some Maplin 0.6w metal 1% resistors which I could use instead or I could buy Vishay individuals a bit cheaper... Do you think this is likely to effect performance?
The driver transistors, do these need screwing down to the heatsink or not?
The main output transistors. Should I use stainless, non magnetic screws? Also should spring washers be used or not? Do you spec a Torque for the screws?
The 5A fuse I guess these be type F ?
Thanks again for all your help. Phono stage working well and sounds great.
Looking at he BOM, some parts I have others I need to buy. I bought a large heatsink on Ebay 370x50x280 for £40. Chopped it in half using a Tile Saw. Looks Big
. That bits sorted !For a start I want to build the Amplifiers up and try them using parts I have at hand for the PSU's. Then I would spend money on nice cases etc..
Wonder if you could advise and answer a few questions for me please, regarding your deign.
I intend to build mono-blocks. I have four 10000uf caps. Would 10000uf per rail be enough? As the current draw would be half that of a single PSU?
I do have a pair of 200VA 19-9-19 Toloroids which may get me away to try the amp (Not screened so not ideal). But again lower magnetic fields as half the current. I did read where larger Toloroids are preferred. Apparently better for reducing the transfer of noise and RFI. So maybe swop out later..
I see you use Bridges rather than high current Hyper-Fast switching Diodes for the PSU. I have some of these so maybe I will use those instead and but snubber caps across them. Again some say bridges are better though?
I guess the 20A bridge earth lifter needs to be 20Amp to withstand mains shorts to earth?
The resistors you specified are expensive as you need to buy 100 off. I have some Maplin 0.6w metal 1% resistors which I could use instead or I could buy Vishay individuals a bit cheaper... Do you think this is likely to effect performance?
The driver transistors, do these need screwing down to the heatsink or not?
The main output transistors. Should I use stainless, non magnetic screws? Also should spring washers be used or not? Do you spec a Torque for the screws?
The 5A fuse I guess these be type F ?
Thanks again for all your help. Phono stage working well and sounds great.
Hello Terps, thanks for your interest and kind feedback on the phono stage. Here are the answers to your KX2 questions
1. 10 000uF per amplifier rail should be ok
2. Using two separate 200 VA toroids will be fine. See the ‘Ground Loops’ presentation in the link below for how to rotate them for minimum noise
3. Hyper fast rectifier diodes will work fine with the amp. Snubbers are always recommended- see Mark Johnson’s thread on this for more details (I think it’s in the Power Supply forum)
4. I specify 20+ Amp bridge rectifier for the ground lifter primarily for safety. Under worst case conditions (mains short to the secondary or amplifier side of the transformer) it must take the full short current until the GFI (aka RCD in some countries) trips. This can be as high as 100A, so the peak repetitive current handling ability of the bridge used is important. I used two back to back diodes in the X-Altra MC/MM preamp. They have a peak current handling ability of 200 A. I always recommend that you try to avoid using a GL if you can, and ground the 0V to the chassis for the best safety. Use a GL only if you have to.
5. Resistors. You will be ok with the ones you have. (Mouser offer a superb range of resistors. I usually buy Xircon which come in at a few pennies a piece, but may be more expensive now because of shortages.)
6. I would screw the drivers down on the heatsink, but it’s not critical for regulating the output bias current, since it’s an active feedback scheme. However, if you do mount them differently, you may have to trim the bias up by changing some resistor values in the Iq control circuit. For simplicity sake, I’d just stick with the existing approach ie screw them down on the heatsink.
7. The screws will not affect the sound - you can use either type safely
8. You must use T fuses - F fuses will pop too easily. If you short the amplifier output, the T fuses will pop. Because the OPS transistors are over specified, they handle it very well. I’ve popped fuses like this quite a few times on the class A amps I’ve designed using the specified devices and you just replace them and you’re good to go again.
1. 10 000uF per amplifier rail should be ok
2. Using two separate 200 VA toroids will be fine. See the ‘Ground Loops’ presentation in the link below for how to rotate them for minimum noise
3. Hyper fast rectifier diodes will work fine with the amp. Snubbers are always recommended- see Mark Johnson’s thread on this for more details (I think it’s in the Power Supply forum)
4. I specify 20+ Amp bridge rectifier for the ground lifter primarily for safety. Under worst case conditions (mains short to the secondary or amplifier side of the transformer) it must take the full short current until the GFI (aka RCD in some countries) trips. This can be as high as 100A, so the peak repetitive current handling ability of the bridge used is important. I used two back to back diodes in the X-Altra MC/MM preamp. They have a peak current handling ability of 200 A. I always recommend that you try to avoid using a GL if you can, and ground the 0V to the chassis for the best safety. Use a GL only if you have to.
5. Resistors. You will be ok with the ones you have. (Mouser offer a superb range of resistors. I usually buy Xircon which come in at a few pennies a piece, but may be more expensive now because of shortages.)
6. I would screw the drivers down on the heatsink, but it’s not critical for regulating the output bias current, since it’s an active feedback scheme. However, if you do mount them differently, you may have to trim the bias up by changing some resistor values in the Iq control circuit. For simplicity sake, I’d just stick with the existing approach ie screw them down on the heatsink.
7. The screws will not affect the sound - you can use either type safely
8. You must use T fuses - F fuses will pop too easily. If you short the amplifier output, the T fuses will pop. Because the OPS transistors are over specified, they handle it very well. I’ve popped fuses like this quite a few times on the class A amps I’ve designed using the specified devices and you just replace them and you’re good to go again.
Last edited:
Struggling to find transistors from Mouser etc.. So I'm looking through the various ones I have.
I have two closely matched pairs of BC547B/557B gain of 270-280 @ 2.5ma. I'm thinking these would be ok for Q1,2?
I have four pairs of 2SC2240BL / 2SA970BL low noise Toshiba devices that have gains of 360 - 400. These I can pair up to 5%. Could you take a look at the spec for these, and tell me if they would be good replacements for Q12,14,13,15 ?
KSA1220A appears to obsolete, and cannot find stock at the usual suppliers. Is there a suitable replacement for the KSC2690A & KSA1220A pair do you know?
Thanks Andy..
I have two closely matched pairs of BC547B/557B gain of 270-280 @ 2.5ma. I'm thinking these would be ok for Q1,2?
I have four pairs of 2SC2240BL / 2SA970BL low noise Toshiba devices that have gains of 360 - 400. These I can pair up to 5%. Could you take a look at the spec for these, and tell me if they would be good replacements for Q12,14,13,15 ?
KSA1220A appears to obsolete, and cannot find stock at the usual suppliers. Is there a suitable replacement for the KSC2690A & KSA1220A pair do you know?
Thanks Andy..
I am a bit distracted at the minute with quite a few non-audio things on my plate, but what I will do in the next week or two is build a kx2 using BD139/140 and check it out.
I am reluctant to say ‘just go ahead and do iypt’ without checking it out first because the originally specified devices that people are struggling to get were/are very good. Very flat hFE vs Ic curve and the gain groupings are tightly grouped. Unfortunately the 140/139 are not as good and the fT is also not the same which may necessitate compensation changes.
Of you can get the original parts then please use these in the meantime.
For the small signal devices, you can get BC557/547 from Mouser. These are not ON devices, but if from the same manufacturer will be ok.
I am reluctant to say ‘just go ahead and do iypt’ without checking it out first because the originally specified devices that people are struggling to get were/are very good. Very flat hFE vs Ic curve and the gain groupings are tightly grouped. Unfortunately the 140/139 are not as good and the fT is also not the same which may necessitate compensation changes.
Of you can get the original parts then please use these in the meantime.
For the small signal devices, you can get BC557/547 from Mouser. These are not ON devices, but if from the same manufacturer will be ok.
Sounds like your a busy chap Andy. No need to rush for us retired folk. Its very nice of you to advise us and I'm sure we all appreciate it. I was talking Amps with my elder brother last Sunday. He said the one that he liked the most was the A1, he said the music drew him in. Small world I said. I'm building a modern take of one. So he's keen to here the thing when I finish it.
Anyway I took a hike, and bought some 2690/1220 pairs from France. How legit they are is anybody's guess. (but then we live in a world where you cannot even trust the main distributors so..). I can do some basic checks when I get them. I ordered the onsemi 546 & 556 from Mouser the 556 is now out to 1st of Feb though. Like a wally I did not order the other manufactures, I thought about it, then forgot. Showing my 62 years!
I have a pair of BC550C I can use for Q7. But then I noticed it states both on the BOM and circuit diagram BC560C which is PNP type In the Transistor notes, of the document it states BC550C .. I assume it is NPN and BC550 and not BC560?
Thanks again...
Anyway I took a hike, and bought some 2690/1220 pairs from France. How legit they are is anybody's guess. (but then we live in a world where you cannot even trust the main distributors so..). I can do some basic checks when I get them. I ordered the onsemi 546 & 556 from Mouser the 556 is now out to 1st of Feb though. Like a wally I did not order the other manufactures, I thought about it, then forgot. Showing my 62 years!
I have a pair of BC550C I can use for Q7. But then I noticed it states both on the BOM and circuit diagram BC560C which is PNP type
In the Transistor notes, of the document it states BC550C .. I assume it is NPN and BC550 and not BC560?Thanks again...
I managed to get all the parts ok, and have built the amplifiers. But have not cased them up properly yet. Despite that I have powered then up, tested them and had a listen over a few days. I have ordered some bits from Modushop to case them up properly. Surprisingly, I have hardly noticed any mains pickup. No doubt it will be worse after I box it up!
Measurements wise I have a couple of possible anomalies but most are close to yours. I didn't need to tweak the offset far from the pot mid position. So that's good. My transformers are both 200VA 19.7-0-19.7 loaded giving -+ 25 Volts DC. Run cool and quiet.
First thing I noticed was the Bias appears to be the reverse of pin 1 & 3. IE pin 3 is Class A not 1. I have on both channels 1.5V over R19 not 1.25. Over both pairs of output resistors I have 305mv Class A, 187mv Class AAB and 64mv Class AB. Which is odd. Not sure if that would be the lower rail volts or gain of Q7. Which are both gain of 505. Standing current is 0.975A Class A. Any advice on this would be appreciated Andy.
I have tried the amp on a pair of Celestion series A speakers which I have not listened to for a while. These are have very clear Mids, but tend towards over bright and edge towards harshness. The sound they created was amazingly good bright but not hard, Jazz Saxophone sounded lifelike. Great mids and tight Bass. Exciting to listen too with lots of detail. I have tried the amplifier with my Tannoy Floor standers. Surprised at the tightness of the bass, extended as well. Expected the clearer mids and imaging. All an improvement over Symasym. Sounded is more balanced than the Celestions, not so bright but imaging superb.
Attached a couple of photos of my lash up..
Measurements wise I have a couple of possible anomalies but most are close to yours. I didn't need to tweak the offset far from the pot mid position. So that's good. My transformers are both 200VA 19.7-0-19.7 loaded giving -+ 25 Volts DC. Run cool and quiet.
First thing I noticed was the Bias appears to be the reverse of pin 1 & 3. IE pin 3 is Class A not 1. I have on both channels 1.5V over R19 not 1.25. Over both pairs of output resistors I have 305mv Class A, 187mv Class AAB and 64mv Class AB. Which is odd. Not sure if that would be the lower rail volts or gain of Q7. Which are both gain of 505. Standing current is 0.975A Class A. Any advice on this would be appreciated Andy.
I have tried the amp on a pair of Celestion series A speakers which I have not listened to for a while. These are have very clear Mids, but tend towards over bright and edge towards harshness. The sound they created was amazingly good bright but not hard, Jazz Saxophone sounded lifelike. Great mids and tight Bass. Exciting to listen too with lots of detail. I have tried the amplifier with my Tannoy Floor standers. Surprised at the tightness of the bass, extended as well. Expected the clearer mids and imaging. All an improvement over Symasym. Sounded is more balanced than the Celestions, not so bright but imaging superb.
Attached a couple of photos of my lash up..
Attachments
Hello Bonsai, first of all, thanks for your generous contributions to the DIY audio community!
I have with great interest read several of your PDF documents detailing the circuits and theory of your designs. I have learned quite a bit from these (although I don't pretend to fully understand it all).
I have been comparing the nx and kx2 schematics, and can see they are very similar circuits, but the kx2 is more developed/refined (with the addtion of eg. EF-VAS).
My question is:
If I modified the kx2 bias network for higher rail voltages (40-45V), wouldn't the kx2 make for a very nice "nx2" amplifer?
- Or would other parts of the circuit also require modification/recalculation?
Note that I'm only thinking of AAB and AB modes here, and would not use Class A mode.
Happy New Year! (soon)
I have with great interest read several of your PDF documents detailing the circuits and theory of your designs. I have learned quite a bit from these (although I don't pretend to fully understand it all).
I have been comparing the nx and kx2 schematics, and can see they are very similar circuits, but the kx2 is more developed/refined (with the addtion of eg. EF-VAS).
My question is:
If I modified the kx2 bias network for higher rail voltages (40-45V), wouldn't the kx2 make for a very nice "nx2" amplifer?
- Or would other parts of the circuit also require modification/recalculation?
Note that I'm only thinking of AAB and AB modes here, and would not use Class A mode.
Happy New Year! (soon)
Terps, apologies I only just saw this post of yours now.
1. The gain of 505 for Q7 is perfect and should not be causing you any problems
2. Your standing currents are a bit on the low side. Can you check that R18 is 220k and then R1 = 18K and R48 33k.
3. Note if you have these resistor positions swapped, the class A and class AB will be swapped on the switch
4. Long shot, please check that C5 is not leaky
5. I assume you are using 1% tolerance resistors for the 1/2 watters but 5% for the 0.33 Ohm 5W ones.
6. Can you conform that your bias stability is good once the amp is warmed up?
Once you have boxed your amp up, provided you keep the supply rail cables tightly twisted and the same with the speaker cables, you should have an amp that is as quiet as your bench top lash-up or better. The noise floor on the kx2 in class A mode at 15W RMS out is very low at about -115 dBr.
Happy to hear you liked the sound!