What? The preamplifier of the amplifier i tried to build?
The preamplifier sounds perfectly fine.
The amplifier i tried to build has too much distortion at high frequencies.
If i raise the volume a lot, it also appears on low frequencies.
This should not happen. This amplifier should have a distortion which is almost unhearable if not totally unhearable.
The preamplifier sounds perfectly fine.
The amplifier i tried to build has too much distortion at high frequencies.
If i raise the volume a lot, it also appears on low frequencies.
This should not happen. This amplifier should have a distortion which is almost unhearable if not totally unhearable.
No, unfortunately.
But still i checked everything anyway.
There are no problems here.
Not in the power supplies, not in the components, not on schematic matching, not on the currents passing through the transistors...
The only thing that remains here is PCB layout.
That is it.
The MJL21193 and MJL21194 are from a company named 'Minos' (Yes, in china...)
But it is still a legit company. I dont think at all that all of this is caused by a different branding of the transistors.
The parameters on the datasheet are very close to the ones of more mainstream brands. The gain is a bit low, but that is the same for the other 'better' brands.
All the rest are made by ON semiconductor.
So, the only thing that remains here is the PCB layout. And it is bad. That is it...
That is what is causing those oscillations and distortions in the amplifier.
I even replaced the 62pF capacitor with higher values but it didnt work, everything i modify makes COMPLETELY no difference.
Those problems persist in the same way no matter what i modify.
Another proof that the problem is the PCB layout.
So due to this, unfortunately, i cant do anything about it.
I could re-make the boards but it seems like i'm not very good at designing PCBs with good layout.
Another option could be that the schematic is still incomplete even with the additional components.
I used MKP capacitors everywhere and the resistors are all ±1% metallized film resistors. (Only one is a metal oxide resistor, and it is the power 62R one).
The 62pF (That now is 470pF but it makes no difference anyway) is the only MLCC class NP0 SMD capacitor.
So the final verdict here is simply:
1) The PCB layout is garbage.
2) The schematic is still incomplete.
And i actually ordered a Nichicon musa non polaeized electrolithic capacitor 100uF 50V and i recieved it quite fast.
And of course, after soldering it in place of the 47uF ceramic... still no difference at all!
But still i checked everything anyway.
There are no problems here.
Not in the power supplies, not in the components, not on schematic matching, not on the currents passing through the transistors...
The only thing that remains here is PCB layout.
That is it.
The MJL21193 and MJL21194 are from a company named 'Minos' (Yes, in china...)
But it is still a legit company. I dont think at all that all of this is caused by a different branding of the transistors.
The parameters on the datasheet are very close to the ones of more mainstream brands. The gain is a bit low, but that is the same for the other 'better' brands.
All the rest are made by ON semiconductor.
So, the only thing that remains here is the PCB layout. And it is bad. That is it...
That is what is causing those oscillations and distortions in the amplifier.
I even replaced the 62pF capacitor with higher values but it didnt work, everything i modify makes COMPLETELY no difference.
Those problems persist in the same way no matter what i modify.
Another proof that the problem is the PCB layout.
So due to this, unfortunately, i cant do anything about it.
I could re-make the boards but it seems like i'm not very good at designing PCBs with good layout.
Another option could be that the schematic is still incomplete even with the additional components.
I used MKP capacitors everywhere and the resistors are all ±1% metallized film resistors. (Only one is a metal oxide resistor, and it is the power 62R one).
The 62pF (That now is 470pF but it makes no difference anyway) is the only MLCC class NP0 SMD capacitor.
So the final verdict here is simply:
1) The PCB layout is garbage.
2) The schematic is still incomplete.
And i actually ordered a Nichicon musa non polaeized electrolithic capacitor 100uF 50V and i recieved it quite fast.
And of course, after soldering it in place of the 47uF ceramic... still no difference at all!
Last edited:
THere are a lot of other factors that may be causing your issues besides the layout. For example, the drivers and input stage aren't decoupled from the output stage. You may have a section of the amp running at a bad operating point, causing clipping. You may have poor phase margin causing it to oscillate. To see what's actually happening you really need an oscilloscope, prefferably an old analog one. Even a cheap USB scope is better than nothing.
But why should it be running bad if all the components match the schematic?
The components themselves are good, the transistors have very close gain. I tested them before soldering to get the best result.
I cant get any type of scope for now.
I'm about to build a better amplifier that i know it will actually work.
Thanks everyone for your help.
The components themselves are good, the transistors have very close gain. I tested them before soldering to get the best result.
I cant get any type of scope for now.
I'm about to build a better amplifier that i know it will actually work.
Thanks everyone for your help.
Transistor batches are all different. For example forward voltage of the base (vbe) often changes and the circuit may need to be adjusted to compensate for it. I think Cordell modelled his amplifiers using On Semi transistors. On Semi hasn't made 2N5551 for a few years now so so you likely aren't using the transistor he originally specified for the amp. This all can have an effect on the finished product and needs to be looked at. Any amp needs to be thoroughly tested and often fine tuned upon assembly. There are tools everyone should have when working on amplifiers. A scope is just one of them.
As I suggested from the start, this amp won't likely be a perfect runner, but it can be a good learning experience. As for the distortion, the first thing to look at is if the output (collector) of Q1 is passing a clean signal, or if it's clipping (not as easy as it sounds). The current flow through it determines what headroom is available to pass a sine wave (or any type of signal) to the following stages. Being the input stage is DC coupled to the VAS section, the voltage needs to be correct there to turn on Q8 the correct amount to allow it to pass a signal, while it turns on the base of Q9 the correct amount for it to pass a signal... it's all a big balancing act and everything needs to be correct.
This amp, like most solid state amps use huge amounts of negative feedback to correct the signal as iit is output, which makes them very tricky to look at the signal as it is passed through the amp. Negative feedback will make the signal look distorted even in a perfectly operating amplifier due to it's "correcting' signal altering the input signal.
As I suggested from the start, this amp won't likely be a perfect runner, but it can be a good learning experience. As for the distortion, the first thing to look at is if the output (collector) of Q1 is passing a clean signal, or if it's clipping (not as easy as it sounds). The current flow through it determines what headroom is available to pass a sine wave (or any type of signal) to the following stages. Being the input stage is DC coupled to the VAS section, the voltage needs to be correct there to turn on Q8 the correct amount to allow it to pass a signal, while it turns on the base of Q9 the correct amount for it to pass a signal... it's all a big balancing act and everything needs to be correct.
This amp, like most solid state amps use huge amounts of negative feedback to correct the signal as iit is output, which makes them very tricky to look at the signal as it is passed through the amp. Negative feedback will make the signal look distorted even in a perfectly operating amplifier due to it's "correcting' signal altering the input signal.
My transistors are ON semiconductor. Months ago i've found a batch of them on ebay. They were new old stock. Only the power ones are not ONsemi.
Calibrating this amplifier does not seem easy.
But with an oscilloscope i probably could have done something.
But i cant have it sooo...
Maybe calibrating a-la-trial and error style?
Changing some resistors maybe?
Also, that negative feedback should set the gain of the amplifier.
The lower the gain the lower the distorsion, i know.
I could decrease the gain but it may still not work
Calibrating this amplifier does not seem easy.
But with an oscilloscope i probably could have done something.
But i cant have it sooo...
Maybe calibrating a-la-trial and error style?
Changing some resistors maybe?
Also, that negative feedback should set the gain of the amplifier.
The lower the gain the lower the distorsion, i know.
I could decrease the gain but it may still not work
You say that your design has distorted output and the frequency response is not flat but how did you determine this?
On the BC-1 that me and Bob designed, I used all the parts from Mouser. We used 2n5551 from Rectron. Diotech is another option. Now that look at Mouser onsemi is still making 2n5551 and offering a few options, that and 2n5401.
How about take your schematic based on your pcb and run it in LTSpice as a means of verification? I hope you read the whole book before you decided to try making a pcb.
On the BC-1 that me and Bob designed, I used all the parts from Mouser. We used 2n5551 from Rectron. Diotech is another option. Now that look at Mouser onsemi is still making 2n5551 and offering a few options, that and 2n5401.
How about take your schematic based on your pcb and run it in LTSpice as a means of verification? I hope you read the whole book before you decided to try making a pcb.
Last edited:
I listened to the speaker.
Low frequencies distortion is lower while i can clearly hear high frequencies distorting and clipping very badly.
Something is wrong, but without a scope its almost impossible to fix.
The distortion should actually be unhearable, both at high and low frequencies.
Low frequencies distortion is lower while i can clearly hear high frequencies distorting and clipping very badly.
Something is wrong, but without a scope its almost impossible to fix.
The distortion should actually be unhearable, both at high and low frequencies.
Well then something is out of sorts for sure. Exactly what is the big ? Did you try LTSpice?
Bc-1 was designed so that you could optional take the feedback from the drivers on the amp front end as a means to test it out, not that your are going to hear anything because the ops is out of the circuit.
Bc-1 was designed so that you could optional take the feedback from the drivers on the amp front end as a means to test it out, not that your are going to hear anything because the ops is out of the circuit.
I drew up the amp in LTSpice. I didn't have models for the drivers so I substituted them. It seems to operate okay. I'm using the new version of LTSpice and haven't figured out how to check phase margin in it yet.
I was just checking that, see attached. Bob's models here. Compensation looks rather conservative, I get ULGF = 480kHz, phase margin = 84.7º and gain margin = 28.7dB.
Cheers,
Cabirio
Sorry, some errors corrected (I forgot to change the Vbe multiplier to 2SC3503 and use exactly the same input and output networks). Pretty much the same result though: ULGF = 475kHz, PM = 84º, GM = 28.5dB.
I would just like to say a huge 'Thank you' to Bob and Rick for the design and development work of the BC-1. Putting something like this together involves a huge amount of work on their part. It took me two years to digest what I was looking at and to work out what goes where but finally my BC-1 is tested and working.....it sounds truly sublime and brings poor recordings to life. I have had some hiccups; thinking I needed to grind down a screw thread on a support bracket with a Dremel without bothering to remove the PS board first - paid for that with a damaged terminal block. Diode inserted back to front on a protection board despite having thought that I's checked twice and finally, after testing a protection board and final assembly of the amplifier misplacing by one pin position a DM35DT temperature sensor resulting in the left hand channel not working after putting all together. I have been up all night on a roll realising that the end was in sight - but made a mistake that could have blown a lot of things up - had I misplaced by one pin position the other way, I'm sure that would have been the case.
I have been listening on the newly commissioned BC-1 for over two hours now and it is giving great pleasure. I can only advise to read the words that come with the PCB's and then re-read until you have the process embedded in your mind. I will be adding power indicators and a switch to the front panel - not feeling that the amplifier may be left on without you knowing (there is no hiss).
Many thanks both. Regards. Neil
I have been listening on the newly commissioned BC-1 for over two hours now and it is giving great pleasure. I can only advise to read the words that come with the PCB's and then re-read until you have the process embedded in your mind. I will be adding power indicators and a switch to the front panel - not feeling that the amplifier may be left on without you knowing (there is no hiss).
Many thanks both. Regards. Neil
For OP: I'll admit, I didn't read every post here. Some folks have already pointed out a number of issues. Another one is based on your post below, your amp was oscillating. When you added the series R+C network at the output, that probably helped make it marginally stable. Meaning with the right stimulus, it will start oscillating again. You would likely have seen this oscillation earlier with a scope. You do have access to a scope, right? You'll need it to help troubleshoot and fix this project properly. This will likely be tough with your level of experience, so you will have to be patient and persistent if you want to get to a successful conclusion with this project. Even if your layout is truly error-free from a net-connectivity perspective, how do you know it is a "good" layout from all the other perspectives that must be considered, such as general layout rules, grounding, trace proximities, feedback take-off points, etc. ?
Good luck!
Good luck!