After some problems... and smoke...
I've monted one channel in the chassi.
And everything is working Perfect...
I've been playing music on One channel for 3h now...
One funny thing, or shall we say HiFi thing..
Is that you can't listen on MP3 in bad quality anymore...
IT SOUNDS CRAP... I thought there were a problem with the amplifier...
Soon as I played a normal CD it sounds great...
So the details is great..
I've never heard it so clear before as now.
There is NO Way I can stand listen to those MP3's again..
Will mount the other channel next days...
I've monted one channel in the chassi.
And everything is working Perfect...
I've been playing music on One channel for 3h now...
One funny thing, or shall we say HiFi thing..
Is that you can't listen on MP3 in bad quality anymore...
IT SOUNDS CRAP... I thought there were a problem with the amplifier...
Soon as I played a normal CD it sounds great...
So the details is great..
I've never heard it so clear before as now.
There is NO Way I can stand listen to those MP3's again..
Will mount the other channel next days...
There are some questions I have,
When setting bias:
What is the correct way ?
No signal and no speaker ?
When setting DC offset:
It's no problem to get 0mv...
But If I measure are other side on the pcb board I then get ~20mv...
Shall I set DC so I get same reading on both sides ?
Bias:
When starting up the bias will first go up, and after 5-10 min it will come down again.. Guess this is normal.
I also should wait 1-2h before setting the final bias...
This is what I did...
But now I've been playing 3h and the bias is comming down... 1mv a time..
I set the bias at: 625mv... then it goes upp to ~660... then back down to ~625... but after 3h, it's down to 599mv...
Shall I wait longer before setting the bias ?!
Guess this is the last things I need to know to get everything perfect...
When setting bias:
What is the correct way ?
No signal and no speaker ?
When setting DC offset:
It's no problem to get 0mv...
But If I measure are other side on the pcb board I then get ~20mv...
Shall I set DC so I get same reading on both sides ?
Bias:
When starting up the bias will first go up, and after 5-10 min it will come down again.. Guess this is normal.
I also should wait 1-2h before setting the final bias...
This is what I did...
But now I've been playing 3h and the bias is comming down... 1mv a time..
I set the bias at: 625mv... then it goes upp to ~660... then back down to ~625... but after 3h, it's down to 599mv...
Shall I wait longer before setting the bias ?!
Guess this is the last things I need to know to get everything perfect...
The small changes in bias are of no consequence.
Big changes that keep moving in the same direction are of much more concern. It could indicate that the thermal stability of the output stage is suspect.
You have no concern here, the readings look fine.
Set the bias and the offset with the input shorted and the output with no external load.
Check the offset does not change when you connect your source in lieu of the input shorting link. If it changes significantly then a redesign is in order.
I found the circuit is very sensitive to the addition of tiny caps on the output terminals. Adding 100nF instantly blew both F5A supply fuses.
The only way to solve this was to increase the Cdom value. Then it sounded terrible, no detail, no sparkle, =not acceptable. I have been waiting almost two years for someone to catch up and assemble another Klone in the hope they can find me a solution. In the meantime my mono is sitting in the corner of the workshop.
Big changes that keep moving in the same direction are of much more concern. It could indicate that the thermal stability of the output stage is suspect.
You have no concern here, the readings look fine.
Set the bias and the offset with the input shorted and the output with no external load.
Check the offset does not change when you connect your source in lieu of the input shorting link. If it changes significantly then a redesign is in order.
I found the circuit is very sensitive to the addition of tiny caps on the output terminals. Adding 100nF instantly blew both F5A supply fuses.
The only way to solve this was to increase the Cdom value. Then it sounded terrible, no detail, no sparkle, =not acceptable. I have been waiting almost two years for someone to catch up and assemble another Klone in the hope they can find me a solution. In the meantime my mono is sitting in the corner of the workshop.
Sensitive to what ?
Damage to caps when connecting speaker ?
- - -
About Bias..
Yesterday I found out that after 4h the bias got stable, +/- 2-3mv
Guess it need 3-4h before getting real stable
OK... But where on the cirrcuit shall I measure...
DC Offset: Signal In (shorted) and depending on where on my design I measure, ( Can measure on 4 TP's I get different readings... ) On Output NPN/PNP side...
I've measured on all points and have now set it to 10mv.
If I set it to 0mv on Output on PNP side, same side where I have my cable to the Outputterminal I get ~25mv on the other NPN side...
Will check if the DC Changes when connecting source.
BIAS:
I've measured on each emittor resistor for similar readings.
Approx: .625, .630. .650 .630
I used resistor where speakar cable is connected as reference on both channels.
Will post pictures if needed....
More pictures will come tonight on rest...
adding a small value capacitor to the speaker output terminal blew the PSU fuses when the amp is idling with no input signal and the standard Cdom is fitted.Fix said:
measure output offset at the speaker output terminal.Fix said:
There can be only one speaker output terminal on a monoblock amplifier, not four.
Strange... I have added a Resistor + Cap as shown in the schematic without any problems at all. ( See pictures few post down... )
I pretty sure that I've started the amp serveral times without any signal / speaker also without any cables at all.
Also I've been connecting Signal cables + speaker without turning off the amp...
I had a problem before where the amp did blow the fuse, soon as I connected signalcable or my oscilioscope..
Problem was that I have a shortage between the Neg - Voltage and Ground... and after 2-3 fuses my 10ohm resistor to ground blew..
Then it did work...
I've fixed that right now, and found out that the 10ohm ground resisitor were bad...
Maybe the Resistor/cap on speaker terminal also burnt ???
Havn't checked that, will do that tonight...
Else I can find any problems...
- - -
Look at post 2313..
Orange cable is Output..
Same on other side..
Yes I have 1 cable as output... but if I measure front and back on that side I get different readings..
Totalt 4 TP's ...
I can get 0mv on one side, and 22mv on the other side.
I've set DC to 10mv, then I have 10mv on each TP.
What is Cdom
a series pair of resistor+ cap is the output Zobel.
Add a capacitor (alone) across the output terminals, i.e. speaker output to audio ground. Try different values from 100pF to 1uF.
I could get away with 68nF, but the scope showed a very significant change in the wave shape. 100nF blew the fuses.
If I excluded the output transistors from the feedback loop. i.e. feedback from drivers only, the circuit is much more tolerant of output loading variations.
Which pic in post 2313?
Cdom (dominant pole capacitor) is the base to collector capacitor, sometimes referred to as Miller compensation capacitor, that tunes the HF response and thus stability of the VAS stage.
Add a capacitor (alone) across the output terminals, i.e. speaker output to audio ground. Try different values from 100pF to 1uF.
I could get away with 68nF, but the scope showed a very significant change in the wave shape. 100nF blew the fuses.
If I excluded the output transistors from the feedback loop. i.e. feedback from drivers only, the circuit is much more tolerant of output loading variations.
Which pic in post 2313?
Cdom (dominant pole capacitor) is the base to collector capacitor, sometimes referred to as Miller compensation capacitor, that tunes the HF response and thus stability of the VAS stage.
Is it needed ?
Will sound be better or protection....
I've tried to drive my test loudspeaker 4ohm, into 4hz without any problem.. The emittor voltage (Bias) went crazy 350mv-750mv (625mv normal..) Looked funny on the speaker...
And my DC Sence PCB didn't brake so I guess there were no problem..
About the cap you talk about, is it used in the original design ?
Shall I remove the Zobel Filter, and add Cap instead ?
- - -
The two last pictures you can see the Orange cable (Output)
It's connected from Front - Back PNP side.. and on NPN side the same.
The Output is connected on both sides...
Depending were I measure I get different readings..
DC Offset: If I measure signal-in to output PNP side, Front I can get 0mv, if I change to back or the NPN side I get 20mv..
I have connected a Speaker cable as shown on the original PNP side...
After many many hours of tears, blood and guts... and some smoke...
There is light in the end of the tunnel.
Another Krell KSA-100 Clone is born.
Been playing for the last 3h with both channels everything mounted in the case.
Before testing the amp with my Infinity IRS Sigma speaker I will test the amp ten more hours so everything holds up.
Some fun facts:
Total weight 37kg
Power consumtion: 550-580w @ 2.49-2.69 Amp / 230V
Bias: ~625mv ( depending on outside temp )
And here is another batch of pictures...
First channel mounted... Lots of space left...
Other channel on it's way...
Both channels in place..
Many thanks to all here at the forum for all help.
I will try to write down everything I've discovered when building the Krell Clone.. So others can start building..
There is light in the end of the tunnel.
Another Krell KSA-100 Clone is born.
Been playing for the last 3h with both channels everything mounted in the case.
Before testing the amp with my Infinity IRS Sigma speaker I will test the amp ten more hours so everything holds up.
Some fun facts:
Total weight 37kg
Power consumtion: 550-580w @ 2.49-2.69 Amp / 230V
Bias: ~625mv ( depending on outside temp )
And here is another batch of pictures...
First channel mounted... Lots of space left...
Other channel on it's way...
Both channels in place..
Many thanks to all here at the forum for all help.
I will try to write down everything I've discovered when building the Krell Clone.. So others can start building..
Thanks....
Some questions:
Will test run the amp some more hours before testing on my real speakers...
But first impressions are: Good !
I can hear if I play a MP3 or CD without any problem.
So the details is great...
Soon as I've tested it amp with my Infinity IRS Simga I'll get back.
RAILS: ~55v
HEATSINKS ~50 Degree Celsius.
I used PAPST 4800X 110v fans, running at full speed.
In future I will test if I can lower the spin, using a 250ohm/50w resistor.. First test shows that this amp need Much cooling as possible.
After 3-4h the temprature in my room is quite high and then the heatsinks get 1-3c warmer and then the BIAS drops drop to ~585mv If I open the window, and some cold air gets in then the bias slowly raises again, and heatsinks get cooled down to 47c.
So I will most likely run fans att full speed...
It's been a long way, but it's been very fun and I've learnt many things. I also documented most work I've done so Heatsinks, PCB's, Softstart, DC Sence etc is easy to make again.
Also my goal were to build as close I could to the original Krell amp in design. And I think I'm pretty close. Hopefully in sound quality too...
For the moment I'm use my laptop as a CD Player... And 2 crappy loudspeakers...
When all tests and I'm sure that the amp don't break down rest of my system is:
Speakers: Infinity IRS Sigma
CD Player: Denon 2560
Pre-Amp/AMP: Denon PMA 1560 (will use pre-out...)
Turntable: Garrard 401 + Denon DL-103
Will buy a new Pre-amp later...
Some questions:
I just only tested the amp with my computer speakers
Will test run the amp some more hours before testing on my real speakers...
But first impressions are: Good !
I can hear if I play a MP3 or CD without any problem.
So the details is great...
Soon as I've tested it amp with my Infinity IRS Simga I'll get back.
BIAS: ~625mv
RAILS: ~55v
HEATSINKS ~50 Degree Celsius.
I used PAPST 4800X 110v fans, running at full speed.
In future I will test if I can lower the spin, using a 250ohm/50w resistor.. First test shows that this amp need Much cooling as possible.
After 3-4h the temprature in my room is quite high and then the heatsinks get 1-3c warmer and then the BIAS drops drop to ~585mv If I open the window, and some cold air gets in then the bias slowly raises again, and heatsinks get cooled down to 47c.
So I will most likely run fans att full speed...
YES...
It's been a long way, but it's been very fun and I've learnt many things. I also documented most work I've done so Heatsinks, PCB's, Softstart, DC Sence etc is easy to make again.
Also my goal were to build as close I could to the original Krell amp in design. And I think I'm pretty close. Hopefully in sound quality too...
For the moment I'm use my laptop as a CD Player... And 2 crappy loudspeakers...
When all tests and I'm sure that the amp don't break down rest of my system is:
Speakers: Infinity IRS Sigma
CD Player: Denon 2560
Pre-Amp/AMP: Denon PMA 1560 (will use pre-out...)
Turntable: Garrard 401 + Denon DL-103
Will buy a new Pre-amp later...
"Will buy a new Pre-amp later..."
Don't buy, build... The Lightspeed would be a snap for you....
http://www.diyaudio.com/forums/showthread.php?postid=924390#post924390
Don't buy, build... The Lightspeed would be a snap for you....
http://www.diyaudio.com/forums/showthread.php?postid=924390#post924390
First real impression...
I couldn't wait any longer to hook up the amp to my real hifi-speakers.
I had tested the amp over 10h with my crappy computerspeakers and had no problems.. I've crossed my fingers and checked so everything looked ok once more, before turning on the power...
First CD, then Pre-amp...
And finally the magic powerswitch loctated rear of my homebuilt amp...
SHOWTIME....
The blue leds were bright blue... I got power... The fans started to spin up... 3 seconds later the Softstart relay did click... and a few seconds after that both DC Sence Relays did click...
No smoke....
No strange buzz,humm or other disturbing noice from speakers...
I turned up the volume just to make sure... perfect...
Time for music....
Just started some cd, and wait until the heatsinks got the right temp....
I have so much impressions, so I don't know where to begin.
I've been playing for two nights now....
"Magic" , "WOW" , "Power" , "Details" , "Deep bass" , "Airness" etc
This Krell Clone rocks... I'm more than happy... Way better than I thought...
Bass:
If I had to turn up the Bass Controller 30% before, I now have everything at Zero... It sounds like I have pressed the Loudness button... Nice warm, deep, bass... Tight and Hard... NOT too much, blurry...
Mid/High: The sound is crisp... and extremly open...
3D !!!
I've played lots of CD's
Kraftwerk: Radioactivity Album
* You can hear everything... They used a BAD mixerboard..., Groundloop problems etc... Fantastic... I've Never heard that before.. If I did, it was Sure not that clear before... And I've listen to this CD many many times before...
Madonna, Metallica, Chemical Brothers etc etc etc....
I can go on and on and on....
This is a great DIY Project... with a fantastic result.
So whats the downside...
Heat... and Fan noice...
Since this is a Real Class-A it will be Hot as hell.
I have fans running at full speed, and the heatsinks are about 50-52c, roomtemp: 23-25c
This is OK... and the bias is stable at 615mv
But the fan noice is too much...
I will start testing to lower the speed...
Also when I close the cover the heat raises 5-10c inside the amp and the bias drops down to ~525mv... and then the sound is not good...
There are lots of solutions, give me some days...
Worst case:
Cut up the cover at back and it will work perfect, but it will not look nice.
I couldn't wait any longer to hook up the amp to my real hifi-speakers.
I had tested the amp over 10h with my crappy computerspeakers and had no problems.. I've crossed my fingers and checked so everything looked ok once more, before turning on the power...
First CD, then Pre-amp...
And finally the magic powerswitch loctated rear of my homebuilt amp...
SHOWTIME....
The blue leds were bright blue... I got power... The fans started to spin up... 3 seconds later the Softstart relay did click... and a few seconds after that both DC Sence Relays did click...
No smoke....
No strange buzz,humm or other disturbing noice from speakers...
I turned up the volume just to make sure... perfect...
Time for music....
Just started some cd, and wait until the heatsinks got the right temp....
I have so much impressions, so I don't know where to begin.
I've been playing for two nights now....
"Magic" , "WOW" , "Power" , "Details" , "Deep bass" , "Airness" etc
This Krell Clone rocks... I'm more than happy... Way better than I thought...
Bass:
If I had to turn up the Bass Controller 30% before, I now have everything at Zero... It sounds like I have pressed the Loudness button... Nice warm, deep, bass... Tight and Hard... NOT too much, blurry...
Mid/High: The sound is crisp... and extremly open...
3D !!!
I've played lots of CD's
Kraftwerk: Radioactivity Album
* You can hear everything... They used a BAD mixerboard..., Groundloop problems etc... Fantastic... I've Never heard that before.. If I did, it was Sure not that clear before... And I've listen to this CD many many times before...
Madonna, Metallica, Chemical Brothers etc etc etc....
I can go on and on and on....
This is a great DIY Project... with a fantastic result.
So whats the downside...
Heat... and Fan noice...
Since this is a Real Class-A it will be Hot as hell.
I have fans running at full speed, and the heatsinks are about 50-52c, roomtemp: 23-25c
This is OK... and the bias is stable at 615mv
But the fan noice is too much...
I will start testing to lower the speed...
Also when I close the cover the heat raises 5-10c inside the amp and the bias drops down to ~525mv... and then the sound is not good...
There are lots of solutions, give me some days...
Worst case:
Cut up the cover at back and it will work perfect, but it will not look nice.
Fix
I suggest that you lower the class A output to 75 watts especially as you are using realtively small speakers. At this stage you don't really need full 100 watts class A. The amp will also run cooler. I have found that when using small speakers that a 30 watt class Aleph 30 meets most of my needs. I don't know how practical it is but it would be nice to have a switch like some Classe amps to switch between 50 or 100 watts A output?
Personally I think the heatsinks you intend to sell needs to be 1-2 centimeters longer than the originals. If one is in position to make the camp run cooler why not implement it ?
The fans will run quiter if they have unrestricted air-flow and are rubber mounted. Consider cutting a round hole in the bottom of the amp for each heat sink. I suspect that your top plate is to close to the heatsink. An air-condition grid might be a better option.
Congratulations- it always nice to sit back and enjoy the end result.
Jozua
I suggest that you lower the class A output to 75 watts especially as you are using realtively small speakers. At this stage you don't really need full 100 watts class A. The amp will also run cooler. I have found that when using small speakers that a 30 watt class Aleph 30 meets most of my needs. I don't know how practical it is but it would be nice to have a switch like some Classe amps to switch between 50 or 100 watts A output?
Personally I think the heatsinks you intend to sell needs to be 1-2 centimeters longer than the originals. If one is in position to make the camp run cooler why not implement it ?
The fans will run quiter if they have unrestricted air-flow and are rubber mounted. Consider cutting a round hole in the bottom of the amp for each heat sink. I suspect that your top plate is to close to the heatsink. An air-condition grid might be a better option.
Congratulations- it always nice to sit back and enjoy the end result.
Jozua
Thanks Jozua and Harry3,
My speaker is quite large... Infinity IRS Sigma...
And they do need Lots of current to play...
When I play music with deep bass, the bias jumps up'n down...
The amp needs to work hard..., bias jumps from +-100mv from 600mv ( in extreme case, else it's superstable...)
Heatsinks:
I've tried to make them close as the original, sure I can make them 1-2 longer. Maybe I should had done that before.
Fans:
I've mounted them with rubber mounts, cut hole in bottom....
I think that I have too high airflow, and then my top won't let the hot air out enough.
If I lower the fan speed, It might get 2-3c warmer, but all air can get out.
Yesterday I did try to raise the amp 3cm from the floor.
And when the cover came on the temp. did raise +2c
Will try some more things before cutting the top.
Also it's summer in Sweden and it's quite hot, +23c inside, in Fall/Winter I have +19-20c inside this will help the amp alot.
For the moment I think the biggest problem is the hot air that don't get out...
If I can fix that, I will have a much cooler amp...
My speaker is quite large... Infinity IRS Sigma...
And they do need Lots of current to play...
When I play music with deep bass, the bias jumps up'n down...
The amp needs to work hard..., bias jumps from +-100mv from 600mv ( in extreme case, else it's superstable...)
Heatsinks:
I've tried to make them close as the original, sure I can make them 1-2 longer. Maybe I should had done that before.
Fans:
I've mounted them with rubber mounts, cut hole in bottom....
I think that I have too high airflow, and then my top won't let the hot air out enough.
If I lower the fan speed, It might get 2-3c warmer, but all air can get out.
Yesterday I did try to raise the amp 3cm from the floor.
And when the cover came on the temp. did raise +2c
Will try some more things before cutting the top.
Also it's summer in Sweden and it's quite hot, +23c inside, in Fall/Winter I have +19-20c inside this will help the amp alot.
For the moment I think the biggest problem is the hot air that don't get out...
If I can fix that, I will have a much cooler amp...
How do I do that...
When I look at the MK2 pictures, there are 2x Yellowcaps / parallell with the big blue caps.
They was not included in the schematic, anyone knows what specs they have ?
Will it help/what do they do...
Infinity is known to make speaker that draws lots of current in strange impedance.. So when that 12" IMG Woofer starts to rock, guess it's hard to drive it.
The bias is changing because the temperature of the Vbe multiplier is changing when the top goes on. Blow on the Vbe multiplier and you'll see the bias change.
Hot amps require lots of airflow - you cannot have too much. It's not clear from the pictures so far what your exhaust arrangement is. Fan noise will be reduced at the same speed/flow if you open up the exhaust path so that the hot air exits more freely. It sounds like your fan is working hard trying to squeeze air out through a restrictive path.
Edit: You'll also get better cooling and less noise if you open up the intake. The slots in the bottom aren't open enough. Cover up 3/4 of your vacuum cleaner nozzle and listen to the fan scream louder, and you'll get an idea what is happening here.
Hot amps require lots of airflow - you cannot have too much. It's not clear from the pictures so far what your exhaust arrangement is. Fan noise will be reduced at the same speed/flow if you open up the exhaust path so that the hot air exits more freely. It sounds like your fan is working hard trying to squeeze air out through a restrictive path.
Edit: You'll also get better cooling and less noise if you open up the intake. The slots in the bottom aren't open enough. Cover up 3/4 of your vacuum cleaner nozzle and listen to the fan scream louder, and you'll get an idea what is happening here.