I was thinking to add the complete driver + 2 Pairs of Power Trannies so the complete the same OP section with all related pasives, practically to be doubled, this way the same conditions established as for the first OPs pair, so even in bridged mode much easier to chose V DC from + - 35V Up
Cheers.
Andreas
Hi,
paralleling always includes the risc of oscillation, just the degree of probability varies. 😉
In principle You may add as many CFPs as the driver stage can accomodate.
But, regarding the cost of a complete L12-2 or a beefier amp module, will it be worth the hassle?
Is Your Q meant in general or specific?
Because for dookie's requirements +-24V trannies would suffice, he won't need 35+ Voltages.
Keep in mind that raising the supply rails will cost more, as it required a beefier Transformer and Capacitor bank also.
Supplied by a 24V transformer, the amps will run into voltage clipping first when driven into overload.
Driving into 8Ohms load there would still be current reserve and SOA left.
The clipping would be clearly audible and gives the chance to turn down volume to safe levels.
You may then be suffciently safe when adding a DC-protection/turn-on delay at the amps outputs.
One could also add Limiters ahead of the power amps, or build a protection circuit that also monitors the OP-Transistors current.
jauu
Calvin
paralleling always includes the risc of oscillation, just the degree of probability varies. 😉
In principle You may add as many CFPs as the driver stage can accomodate.
But, regarding the cost of a complete L12-2 or a beefier amp module, will it be worth the hassle?
Is Your Q meant in general or specific?
Because for dookie's requirements +-24V trannies would suffice, he won't need 35+ Voltages.
Keep in mind that raising the supply rails will cost more, as it required a beefier Transformer and Capacitor bank also.
Supplied by a 24V transformer, the amps will run into voltage clipping first when driven into overload.
Driving into 8Ohms load there would still be current reserve and SOA left.
The clipping would be clearly audible and gives the chance to turn down volume to safe levels.
You may then be suffciently safe when adding a DC-protection/turn-on delay at the amps outputs.
One could also add Limiters ahead of the power amps, or build a protection circuit that also monitors the OP-Transistors current.
jauu
Calvin
type of solder
These boards come with a solder coating as usual. Is this lead free solder ? I need to turn up the temp quite a bit to get the solder to melt on the board . Melts at a temp higher than 60/40 solder. I guess all boards are ROHS now ?
Thanks.
These boards come with a solder coating as usual. Is this lead free solder ? I need to turn up the temp quite a bit to get the solder to melt on the board . Melts at a temp higher than 60/40 solder. I guess all boards are ROHS now ?
Thanks.
I would expect all new PCB to be rohs compliant and thus use a higher melting point solder.
But when you flood on a low temperature solder the original solder becomes very diluted.
This does make the molten solder pool contaminated and will change it's melting/solidifying temp from the eutectic solder you are using (60/40 is not a eutectic). I.e it will solidify at a slightly higher temperature and will be slightly pasty due to the contaminants.
But when you flood on a low temperature solder the original solder becomes very diluted.
This does make the molten solder pool contaminated and will change it's melting/solidifying temp from the eutectic solder you are using (60/40 is not a eutectic). I.e it will solidify at a slightly higher temperature and will be slightly pasty due to the contaminants.
Thanks for that info.
I do use 63/37 also.....at times. The stock is low ! Locally 63/37 is only available against order unless it is Kester on Ebay which is twice as expensive as the local brand. Most likely due to customs duties.
But I did manage to solder everything well ( with 60/40 ) at just above 350 Deg C setting. Not sure if the tip reaches that temp. Solder joints look reasonably shiny.
I do use 63/37 also.....at times. The stock is low ! Locally 63/37 is only available against order unless it is Kester on Ebay which is twice as expensive as the local brand. Most likely due to customs duties.
But I did manage to solder everything well ( with 60/40 ) at just above 350 Deg C setting. Not sure if the tip reaches that temp. Solder joints look reasonably shiny.
test your iron setting.
wind a tiny ring of one turn around the tip of your iron.
set the iron to 180°C and leave it to warm up.
increase the temp setting by 5Cdegrees and leave it a while. Increase in steps until the solder melts.
That gives you a comparison for setting temperature to solder melting temperature.
If your 183 eutectic solder melts at 195°C, then your iron is calibrated very well. 60/40 has a pasty range from 190 to 183°C
If you have to go up to 230°C, then allow for that error whenever you set your iron.
wind a tiny ring of one turn around the tip of your iron.
set the iron to 180°C and leave it to warm up.
increase the temp setting by 5Cdegrees and leave it a while. Increase in steps until the solder melts.
That gives you a comparison for setting temperature to solder melting temperature.
If your 183 eutectic solder melts at 195°C, then your iron is calibrated very well. 60/40 has a pasty range from 190 to 183°C
If you have to go up to 230°C, then allow for that error whenever you set your iron.
Last edited:
How many of you have removed C14 ( 100pF) as recommended by the designer ?
Is the amp stable and does it sound any better ? I've not soldered it on my board . My board will be ready in another week or two. I have just ordered the heat sinks. I am also a bit concerned about the fixed resistors on the bias circuit. I'll use some protection to ensure that it will not emit blue smoke in case the values don't work ! I hope they are OK as I don't fancy de-soldering / re-soldering anything. The tracks and pads are pretty small and I don't want to lift off anything. I had a nightmare removing a TO92 transistor as it was turned round. Luckily nothing lifted off !
I was supplied a Samwah 10uF electrolytic cap for the input. Will that do justice to the amp ? I haven't put it in yet. Been mulling over it. Might just use a larger film cap outside ! I have some time to decide on that. If I do use an Elco, is there any suggestion what brand to use. I can't get Panasonic here. Maybe I should do some testing here.
Is the amp stable and does it sound any better ? I've not soldered it on my board . My board will be ready in another week or two. I have just ordered the heat sinks. I am also a bit concerned about the fixed resistors on the bias circuit. I'll use some protection to ensure that it will not emit blue smoke in case the values don't work ! I hope they are OK as I don't fancy de-soldering / re-soldering anything. The tracks and pads are pretty small and I don't want to lift off anything. I had a nightmare removing a TO92 transistor as it was turned round. Luckily nothing lifted off !
I was supplied a Samwah 10uF electrolytic cap for the input. Will that do justice to the amp ? I haven't put it in yet. Been mulling over it. Might just use a larger film cap outside ! I have some time to decide on that. If I do use an Elco, is there any suggestion what brand to use. I can't get Panasonic here. Maybe I should do some testing here.
Last edited:
Ive just finished a project using these boards and C14 was removed when i got them , looking at the pads it deffo seams to have been removed after initial soldering.
I used a 300va 2x30v tranny for +/-46.2v rails.
Sound wise i think it sounds very very good with some suprising grunt down the low freq end.
One thing it runs really cool on anything ive connected so far which has only been some cheap nasty plastic ariston 8ohm things for 1st tests and allso tried on some proac tablette 8ohm. Both times i did give it some welly for half hr or so and couldnt really generate any more heat than warm to touch on the sinks.
Not that im complaining 🙂
I used a 300va 2x30v tranny for +/-46.2v rails.
Sound wise i think it sounds very very good with some suprising grunt down the low freq end.
One thing it runs really cool on anything ive connected so far which has only been some cheap nasty plastic ariston 8ohm things for 1st tests and allso tried on some proac tablette 8ohm. Both times i did give it some welly for half hr or so and couldnt really generate any more heat than warm to touch on the sinks.
Not that im complaining 🙂
Heat build up.
Basically all class AB amps have a fixed amount of power loss which dissipates as heat . Maximum power loss being at about 1/3rd maximum output of the amp. Temp also being highest there. So the temperature rise would be due to heat transfer from the transistor junction to the air.
Thermal resistance from junction to transistor tab , tab to heat sink ( through insulator) , heat sink to air. Improving each junction would decrease the temperature rise .
So without knowing these factors it isn't possible to compare the temperature rise of different amps. Circuits are not responsible for the heat build up. All AB's will behave the same up to the transistor junction. Rest depends on power transistor type, insulators, heat sink type, ambient conditions etc.
Basically all class AB amps have a fixed amount of power loss which dissipates as heat . Maximum power loss being at about 1/3rd maximum output of the amp. Temp also being highest there. So the temperature rise would be due to heat transfer from the transistor junction to the air.
Thermal resistance from junction to transistor tab , tab to heat sink ( through insulator) , heat sink to air. Improving each junction would decrease the temperature rise .
So without knowing these factors it isn't possible to compare the temperature rise of different amps. Circuits are not responsible for the heat build up. All AB's will behave the same up to the transistor junction. Rest depends on power transistor type, insulators, heat sink type, ambient conditions etc.
Effect of ( cap ) parts changes
I had the C14 cap (100pF disc) on my board as it was a kit. I removed it and I think the treble is a shade crisper. Not too much but just noticeable. I'll do it again later when I have time to be "sure" my mind was not playing tricks. Amp however is stable without it. I have not tried at it at very high volume or tested it with square waves yet !
Next change was the input capacitor which I didn't like even when I saw it ! It was a 10uF/16V Samwah . So the amp was OK with it. Output dc offset 4mV at the speakers.
Then I removed it. As I expected there was a big jump in quality !
All round. Bass more defined, HF cleaner and voice really improved.
Price to pay was a -19mV dc offset at the speakers ( not a problem at all ). Additionally any dc offset at the DAC or source would come through ( possibly not good in some systems!).
In my case I found that my DAC gave a small dc step when skipping tracks . Not much and not very visible on the speakers but you can easily see it on the scope and hear a small "tick" on the speakers.
Now a suggestion : ONLY if you have installed a dc speaker protection relay in your system !
If your source has an output cap on it or very low dc offset then skip the input cap on this amp ! Will bring out the best in the amp.
If you need an input cap or just want to be careful, install a VERY good quality film cap ( which undoubtedly will be large ) off the board . Short the location of the input cap with a wire and fix the input film cap off the board, maybe near the input sockets on your chassis.
Since I tend to be forgetful sometimes I fear I might connect it to something with dc without realising it and so would prefer to install a cap . The best method is to fix a small switch or push button ( externally ) to switch between different caps and try out all you have and determine the best for the location. If you don't want to do this I might just suggest a Solen PP film cap. Maybe anything from 2.2uF ( -3dB at 7.2 Hz ) to 10uF ( -3dB at 1.6 Hz)! Some polyesters are also good. You may need to try them out. Listen very carefully to the mid range , especially on noisy vocal tracks !
Nice amp. Now I will have to build a chassis for it ! I got some nice heat sinks for it !
I had the C14 cap (100pF disc) on my board as it was a kit. I removed it and I think the treble is a shade crisper. Not too much but just noticeable. I'll do it again later when I have time to be "sure" my mind was not playing tricks. Amp however is stable without it. I have not tried at it at very high volume or tested it with square waves yet !
Next change was the input capacitor which I didn't like even when I saw it ! It was a 10uF/16V Samwah . So the amp was OK with it. Output dc offset 4mV at the speakers.
Then I removed it. As I expected there was a big jump in quality !
All round. Bass more defined, HF cleaner and voice really improved.
Price to pay was a -19mV dc offset at the speakers ( not a problem at all ). Additionally any dc offset at the DAC or source would come through ( possibly not good in some systems!).
In my case I found that my DAC gave a small dc step when skipping tracks . Not much and not very visible on the speakers but you can easily see it on the scope and hear a small "tick" on the speakers.
Now a suggestion : ONLY if you have installed a dc speaker protection relay in your system !
If your source has an output cap on it or very low dc offset then skip the input cap on this amp ! Will bring out the best in the amp.
If you need an input cap or just want to be careful, install a VERY good quality film cap ( which undoubtedly will be large ) off the board . Short the location of the input cap with a wire and fix the input film cap off the board, maybe near the input sockets on your chassis.
Since I tend to be forgetful sometimes I fear I might connect it to something with dc without realising it and so would prefer to install a cap . The best method is to fix a small switch or push button ( externally ) to switch between different caps and try out all you have and determine the best for the location. If you don't want to do this I might just suggest a Solen PP film cap. Maybe anything from 2.2uF ( -3dB at 7.2 Hz ) to 10uF ( -3dB at 1.6 Hz)! Some polyesters are also good. You may need to try them out. Listen very carefully to the mid range , especially on noisy vocal tracks !
Nice amp. Now I will have to build a chassis for it ! I got some nice heat sinks for it !
Last edited:
I forgot to add that I added a small heatsink on the driver and bias transistors. Originally they just sat on the pcb . I added an L shaped aluminum heatsink to that so that the heat conductivity between them was better and to keep them cooler than otherwise. I find they get slightly warm while playing music even at low levels. So I think it was a good idea. Without that they would have become much hotter.
The other change which has already been explored is the film cap on the dc decoupling cap on the NFB loop. I don't like the electrolytic cap there either but it might lead to excess dc at the output due to the source offset. I haven't added the film cap there yet.
What needs to be determined is, whether it is better to dc couple the nfb stage or have a direct dc input ( no input cap ). dc decoupling both without an additional opamp to control dc is not good. Or is that solution even better ?
I dislike the idea of chopping up this board for so many experiments!
Better to give it it's own supply and chassis ! It's a very nice amp as it is. Must try to dump the input cap as far as possible !
The other change which has already been explored is the film cap on the dc decoupling cap on the NFB loop. I don't like the electrolytic cap there either but it might lead to excess dc at the output due to the source offset. I haven't added the film cap there yet.
What needs to be determined is, whether it is better to dc couple the nfb stage or have a direct dc input ( no input cap ). dc decoupling both without an additional opamp to control dc is not good. Or is that solution even better ?
I dislike the idea of chopping up this board for so many experiments!
Better to give it it's own supply and chassis ! It's a very nice amp as it is. Must try to dump the input cap as far as possible !
DC coupling of the Power Amplifier is a big risk.
I prefer the safety of DC blocking of both the output from the Source and the NFB feedback.
The extra stages to protect the amp and the speaker is not worth the effort and the resources and the complications to save one from a broken system.
BTW,
a reasonable quality 1uF polypropylene output capacitor feeding a 100k input impedance gets one to a virtually distortionless link between source and receiver.
I prefer the safety of DC blocking of both the output from the Source and the NFB feedback.
The extra stages to protect the amp and the speaker is not worth the effort and the resources and the complications to save one from a broken system.
BTW,
a reasonable quality 1uF polypropylene output capacitor feeding a 100k input impedance gets one to a virtually distortionless link between source and receiver.
Right, dc coupling has it's hazards as I have already pointed out.
However if the source is known with very low dc ( like a cap coupled output stage of a preamp) , a second coupling capacitor can be avoided. Since this is for DIY guys , they can look at all this. Certainly not something that can be recommended to the general public ! 😉
It is certainly safer to always have a dc protection relay at the output of the amp ( or SS relay !).
The dc blocking cap at the NFB is certainly a sensitive area. The dc gain jumps up to the gain of the amp and small offsets at the input can cause havoc.
However in both areas the gain in performance is often audible. This cap can be eliminated in an amp with a JFET input. Good dc balance and offset correction is required. Thermal stability also becomes important .
But everyone should try it sometime even if it is just to know what the difference is. It's all about DIY and that occasional blue smoke ! 🙂
The input impedance here seen by the capacitor is 11 K ohms . So a slightly larger cap would be preferred than 1uF. The 100K is just to ground the other leg of the cap. So Z in ( as seen by the preamp or source ) becomes about 11K //100K.
If we find a very good input cap we can still loose the electrolytic at the NFB point . However we will require offset adjustment and have to check the thermal stability of the offset voltage. If we can get even +/- 50 mV it will be fine.
So there is a lot of tweaking that one can do ! 😉
I forgot to add that anyone trying these should always disconnect the speaker when checking for dc at the output . Even if there is a protection circuit in place. Connect the speaker only after ensuring the dc offset is acceptable.
Better to be safe than sorry. A fuse in the speaker line as protection isn't good enough ! Be careful !
It's like bungee jumping. If something bad happens it can kill you , but still lots of people do it. Here only the amp could get killed ! That's much safer I guess ! 😀
However if the source is known with very low dc ( like a cap coupled output stage of a preamp) , a second coupling capacitor can be avoided. Since this is for DIY guys , they can look at all this. Certainly not something that can be recommended to the general public ! 😉
It is certainly safer to always have a dc protection relay at the output of the amp ( or SS relay !).
The dc blocking cap at the NFB is certainly a sensitive area. The dc gain jumps up to the gain of the amp and small offsets at the input can cause havoc.
However in both areas the gain in performance is often audible. This cap can be eliminated in an amp with a JFET input. Good dc balance and offset correction is required. Thermal stability also becomes important .
But everyone should try it sometime even if it is just to know what the difference is. It's all about DIY and that occasional blue smoke ! 🙂
The input impedance here seen by the capacitor is 11 K ohms . So a slightly larger cap would be preferred than 1uF. The 100K is just to ground the other leg of the cap. So Z in ( as seen by the preamp or source ) becomes about 11K //100K.
If we find a very good input cap we can still loose the electrolytic at the NFB point . However we will require offset adjustment and have to check the thermal stability of the offset voltage. If we can get even +/- 50 mV it will be fine.
So there is a lot of tweaking that one can do ! 😉
I forgot to add that anyone trying these should always disconnect the speaker when checking for dc at the output . Even if there is a protection circuit in place. Connect the speaker only after ensuring the dc offset is acceptable.
Better to be safe than sorry. A fuse in the speaker line as protection isn't good enough ! Be careful !
It's like bungee jumping. If something bad happens it can kill you , but still lots of people do it. Here only the amp could get killed ! That's much safer I guess ! 😀
Last edited:
is a DC blocked input from the Source.
That's one DC route eliminated.
Just the NFB to get DC blocked now.
Input capacitors tested
I tried out several input capacitors. Solen MKP 4.7 uF film cap, Mundorf MCap 4.7uF film cap, Epcos 1.5uF motor run cap, Generic MPP 4uF motor run cap.
All compared with the amp with no input cap. My choice was in the following order. MPP4uF, Solen, Mundorf, Epcos. The MPP4 sounds best. Difference from direct dc input was a very slight softening of bass transient and/or slightly less clear midrange and slight loss of that sheen in HF. Differences are not huge but I think I hear differences. Someone else might not have any preference at all or might have a different rating for them.
The difference between direct dc and the MPP4uF is so small that it's safer to use the cap and you wouldn't hear any diff while playing music. 😉
Since all the capacitors have not been used much they might just get better with time ? Might become as good as a direct dc input connection ?
The capacitor is huge at about 35 mm in diameter and about 60 mm long. So it will have to stay near the input socket. I'll add a push button to bypass the cap when I want to ! It's also the cheapest capacitor in the list above !
This amp is really nice. My bias current is very small but I don't feel it worthwhile meddling with it to increase it. Low level fine HF signals come out super clean. I have not yet checked it out on the scope with 10 Khz at low level. It sounds clean and so I think I will not see anything on the scope to worry me. I don't plan to check the distortion as it 'sounds good' to my ears !
I wont worry about 0.1% and 0.0001%. It means nothing when it comes to what my ears tell me and in this case the amp is 'very good' ! 🙂
I pick up my 300VA transformer on Monday after which I will put the amp into a chassis. The first amp to go into a proper chassis after many years ! 🙂
I tried out several input capacitors. Solen MKP 4.7 uF film cap, Mundorf MCap 4.7uF film cap, Epcos 1.5uF motor run cap, Generic MPP 4uF motor run cap.
All compared with the amp with no input cap. My choice was in the following order. MPP4uF, Solen, Mundorf, Epcos. The MPP4 sounds best. Difference from direct dc input was a very slight softening of bass transient and/or slightly less clear midrange and slight loss of that sheen in HF. Differences are not huge but I think I hear differences. Someone else might not have any preference at all or might have a different rating for them.
The difference between direct dc and the MPP4uF is so small that it's safer to use the cap and you wouldn't hear any diff while playing music. 😉
Since all the capacitors have not been used much they might just get better with time ? Might become as good as a direct dc input connection ?
The capacitor is huge at about 35 mm in diameter and about 60 mm long. So it will have to stay near the input socket. I'll add a push button to bypass the cap when I want to ! It's also the cheapest capacitor in the list above !
This amp is really nice. My bias current is very small but I don't feel it worthwhile meddling with it to increase it. Low level fine HF signals come out super clean. I have not yet checked it out on the scope with 10 Khz at low level. It sounds clean and so I think I will not see anything on the scope to worry me. I don't plan to check the distortion as it 'sounds good' to my ears !
I wont worry about 0.1% and 0.0001%. It means nothing when it comes to what my ears tell me and in this case the amp is 'very good' ! 🙂
I pick up my 300VA transformer on Monday after which I will put the amp into a chassis. The first amp to go into a proper chassis after many years ! 🙂
Last edited:
What impedance was the 1u5F through to 4u7F feeding into?
What was the roll-off frequency of each of the capacitor filters?
Please repeat the test with all the uF matching the highest value of 4u7F.
eg. 680nF||4uF ~ 4u7F, 3u3F||1u5F ~ 4u7F
What was the roll-off frequency of each of the capacitor filters?
Please repeat the test with all the uF matching the highest value of 4u7F.
eg. 680nF||4uF ~ 4u7F, 3u3F||1u5F ~ 4u7F
Just tried a 10uF motor run cap.Very large . 35 x 70 mm.
Very nice. Compared to direct dc connection it is almost as good.As I said earlier, I will have provision to switch it in and out for those really 'special recordings' ! 🙂
10uF with the 10K Zin gives a -3dB at 1.6Hz ! A -3dB at 4 Hz would be good enough and that requires a cap of about 4 uF.
Very nice. Compared to direct dc connection it is almost as good.As I said earlier, I will have provision to switch it in and out for those really 'special recordings' ! 🙂
10uF with the 10K Zin gives a -3dB at 1.6Hz ! A -3dB at 4 Hz would be good enough and that requires a cap of about 4 uF.
Were all your previous tests carried out with the same Rin=10k?
That would almost certainly account for much of the change heard between the 1u5F and 4uF and 4u7F.
They were each rolling off different quantities of the extreme low bass, and adding different quantities of relative phase shift of the low bass and adding different quantities of low frequency distortion.
These three sets of changes, due solely to using different input filter time constants, are likely to be audible !
The input filter Time Constants for a 10k Rin are:
10uF 100ms
4u7F 47ms
4uF 40ms
1u5F 15ms
That would almost certainly account for much of the change heard between the 1u5F and 4uF and 4u7F.
They were each rolling off different quantities of the extreme low bass, and adding different quantities of relative phase shift of the low bass and adding different quantities of low frequency distortion.
These three sets of changes, due solely to using different input filter time constants, are likely to be audible !
The input filter Time Constants for a 10k Rin are:
10uF 100ms
4u7F 47ms
4uF 40ms
1u5F 15ms