Hi Jozua,
How can we make the amplifier less sensitive to cabling and yet maintain or improve the sound quality?
When you talk of fine cored wire, what exactly have you used before and after?
I have 42strand 1.5sqmm, 79strand 2.5sqmm, 412strand 4sqmm and a few others. I wonder if any of these get into the ballpark?
I think the question should be
How can we make the amplifier less sensitive to cabling and yet maintain or improve the sound quality?
When you talk of fine cored wire, what exactly have you used before and after?
I have 42strand 1.5sqmm, 79strand 2.5sqmm, 412strand 4sqmm and a few others. I wonder if any of these get into the ballpark?
Cable choice
AndrewT
Have a look at the Van Den Hul Site and you will see the cable's I am referring to.
Recommendations what to use- as much as I dislike the idea of resoldering the same joints I intend to "test" two more cables. I am open to suggestions as to what to use. Maybe Mark has some ideas based on his 80B experiences?
I am using my 50.2 as a benchmark sound so if my 100.2 it does not at least equal the sound of the 50.2 I know something is wrong. My 100.1 has supertight bass and the 100.2 is a better amp so I can easily determine when the 100.2 is not operating at it's full potential when the bass is not as good as the 100.1.
Mark
Did you see my request for 2 more diodes- please quote if you still have stock.
AndrewT
Have a look at the Van Den Hul Site and you will see the cable's I am referring to.
Recommendations what to use- as much as I dislike the idea of resoldering the same joints I intend to "test" two more cables. I am open to suggestions as to what to use. Maybe Mark has some ideas based on his 80B experiences?
I am using my 50.2 as a benchmark sound so if my 100.2 it does not at least equal the sound of the 50.2 I know something is wrong. My 100.1 has supertight bass and the 100.2 is a better amp so I can easily determine when the 100.2 is not operating at it's full potential when the bass is not as good as the 100.1.
Mark
Did you see my request for 2 more diodes- please quote if you still have stock.
Cables
Hi
Here is a picture of the yellow van Den Hul cable that was added between the output devices. It was done simply as a test and the other side of the output devices cable's were left untouched as the whole heatsink block needs to taken out to gain access to the solder joints.
The Black wires running from the PCB to the output devices are CS18 Van den Hull wire.
After a weekend of listening I switched back to my 50.2 and suddenly the 50.2 sounded thin compared to the rich and full detailed sound of the 100.2. It is incredible to think that such a short piece of wire could have such a great impact on the sound.
Suddenly I am beginning to wonder how my Clone 50 will sound if I change the wiring between the output devices to a thicker multistrand cable?
Jozua
Hi
Here is a picture of the yellow van Den Hul cable that was added between the output devices. It was done simply as a test and the other side of the output devices cable's were left untouched as the whole heatsink block needs to taken out to gain access to the solder joints.
The Black wires running from the PCB to the output devices are CS18 Van den Hull wire.
After a weekend of listening I switched back to my 50.2 and suddenly the 50.2 sounded thin compared to the rich and full detailed sound of the 100.2. It is incredible to think that such a short piece of wire could have such a great impact on the sound.
Suddenly I am beginning to wonder how my Clone 50 will sound if I change the wiring between the output devices to a thicker multistrand cable?
Jozua
Attachments
Hi Jozua,
The more I hear about your findings the more convinced I am that the amp has become cable sensitive BECAUSE the NFB tapping point is in the wrong location.
I don't have the references you luckily have available. Could you do the NFB experiment and report the findings?
It involves removing one end of the NFB resistor from the PCB.
I had my amp dissipating 250W and no problems last night.
Slight concern about variation on Ib across the range of 12 output devices.
My choice of low Re=0r35 seems worse than I expected.
Last night I had mains at 235Vac and open circuit voltage of +-50.7Vdc. When biased to 310mA the PSU has fallen to +-50.3Vdc
By the time I had it up to Iq=2.56A (Ib~=2.5A) the PSU voltage was down to +-48.9Vdc (the Zener currents must be tiny, but not measured yet).
The PSU voltage drop is MUCH higher than I expected. Is this normal for 2.5Amps? or should I start looking for a transformer or rectifier or smoothing capacitance problem?
The more I hear about your findings the more convinced I am that the amp has become cable sensitive BECAUSE the NFB tapping point is in the wrong location.
I don't have the references you luckily have available. Could you do the NFB experiment and report the findings?
It involves removing one end of the NFB resistor from the PCB.
I had my amp dissipating 250W and no problems last night.
Slight concern about variation on Ib across the range of 12 output devices.
My choice of low Re=0r35 seems worse than I expected.
Last night I had mains at 235Vac and open circuit voltage of +-50.7Vdc. When biased to 310mA the PSU has fallen to +-50.3Vdc
By the time I had it up to Iq=2.56A (Ib~=2.5A) the PSU voltage was down to +-48.9Vdc (the Zener currents must be tiny, but not measured yet).
The PSU voltage drop is MUCH higher than I expected. Is this normal for 2.5Amps? or should I start looking for a transformer or rectifier or smoothing capacitance problem?
Andrew,
I fully agree with your NFB theory. The experiment is a bit tedious on the original (I've stripped and rebuilt that one once), so I'd rather do it one the clone project and if successful redo it on the original. I'd also consider a more "star" NFB connection by cutting the thick trace connecting the positive and negative drivers' 75R resistors (the one at the very top of the board), running a thick wire between the NPN and PNP sides instead with a cut in the middle, and connecting both the 75R's, NFB and speaker load at that point. This way it's perfectly symmetrical. I'd avoid taking the NFB at the speaker terminals. It's not too visible on the picture, but the speaker connection is at the left-hand side of the added yellow wire, so it's not only asymmetric w.r.t the transistors, but also the NFB.
That drop is very good compared to what I've seen on other amps. The original Mk2 drops more than 3V IIRC, and it uses massive Avels with some of the best regulation specs of all toroidals.
Regarding Re choice: All the older KSA's, including the 100Mk2, used 1R. However, the 50Mk2 (which I think is slightly younger) used 0R5. It would be interesting to see what the S model used.
I fully agree with your NFB theory. The experiment is a bit tedious on the original (I've stripped and rebuilt that one once), so I'd rather do it one the clone project and if successful redo it on the original. I'd also consider a more "star" NFB connection by cutting the thick trace connecting the positive and negative drivers' 75R resistors (the one at the very top of the board), running a thick wire between the NPN and PNP sides instead with a cut in the middle, and connecting both the 75R's, NFB and speaker load at that point. This way it's perfectly symmetrical. I'd avoid taking the NFB at the speaker terminals. It's not too visible on the picture, but the speaker connection is at the left-hand side of the added yellow wire, so it's not only asymmetric w.r.t the transistors, but also the NFB.
That drop is very good compared to what I've seen on other amps. The original Mk2 drops more than 3V IIRC, and it uses massive Avels with some of the best regulation specs of all toroidals.
Regarding Re choice: All the older KSA's, including the 100Mk2, used 1R. However, the 50Mk2 (which I think is slightly younger) used 0R5. It would be interesting to see what the S model used.
really!
I am not used to this ClassA building, so I'm learning all over again.
Is anyone else upto powering up yet?
With these low voltages I have a choice, either add some windings to my transformers (37Vac) or reduce r11 & r12 to maintain 13mA into the zeners and LTPs.
Which do you recommend?
Cables
AndrewT
I aim waiting for some new (SCS12) Van den Hul cable to arrive before any further changes are implemented. I will discuss your proposal with Pierre.
If my initial results are anything to go by then the correct cable choice could add a significant "WOW" factor to this amp.
Jozua
AndrewT
I aim waiting for some new (SCS12) Van den Hul cable to arrive before any further changes are implemented. I will discuss your proposal with Pierre.
If my initial results are anything to go by then the correct cable choice could add a significant "WOW" factor to this amp.
Jozua
Hi,
I just substituted a 40Vac transformer for the 37Vac and ran the amp at 270W dissipation (110V & 2.46A) for an hour.
It measured a bit too hot for my liking. Sinks sitting at between 65 & 70degC when ambient was 22degC. Derating for deltaT=47Cdeg is obviously considerable.
I think that convinces me to stay with the lower voltage and increase the Zener current back to the 13mA (52V-39V) standard value.
This lower rail voltage of +-49V does have the disadvantage of the Zener dissipation varying more as mains voltage swings. But, ultimately I plan to regulate the front end supplies, so this is not a long term problem.
Does anyone want to convince me that plastic devices run at Tc~=80degC is good for them?
I just substituted a 40Vac transformer for the 37Vac and ran the amp at 270W dissipation (110V & 2.46A) for an hour.
It measured a bit too hot for my liking. Sinks sitting at between 65 & 70degC when ambient was 22degC. Derating for deltaT=47Cdeg is obviously considerable.
I think that convinces me to stay with the lower voltage and increase the Zener current back to the 13mA (52V-39V) standard value.
This lower rail voltage of +-49V does have the disadvantage of the Zener dissipation varying more as mains voltage swings. But, ultimately I plan to regulate the front end supplies, so this is not a long term problem.
Does anyone want to convince me that plastic devices run at Tc~=80degC is good for them?
Re: How does it sound ?
Oh, how I wish you had provided 0.1inch pitch connections onto the PCB. Trying to drill these through the pads that are available is not easy.
Pic, "nice looking", don't kid yourself!
I will be at least days if not weeks away from listening. But I may get to test signals within a short while.Jozua said:
Oh, how I wish you had provided 0.1inch pitch connections onto the PCB. Trying to drill these through the pads that are available is not easy.
Pic, "nice looking", don't kid yourself!
Re: Re: How does it sound ?
No, I am talking about the connections to get onto or off board.
eg, audio input, +V & -V at LTP end, +V & -V at driver end, Driver outputs, across VR2 for dual bias current (winter/summer).
I have inserted some 3way 0.2inch terminal blocks with the middle pin removed to fit the driver outputs, but they are a fiddle to screw up repeatedly when assembling and 2way at the output/PSU V connections and at the input grounds because I could not get 0.1inch to fit.
AndrewT said:
Not for components, you did pretty well there, although some of the electrolytics could have done with dual pitch connections.PWatts said:
No, I am talking about the connections to get onto or off board.
eg, audio input, +V & -V at LTP end, +V & -V at driver end, Driver outputs, across VR2 for dual bias current (winter/summer).
I have inserted some 3way 0.2inch terminal blocks with the middle pin removed to fit the driver outputs, but they are a fiddle to screw up repeatedly when assembling and 2way at the output/PSU V connections and at the input grounds because I could not get 0.1inch to fit.
PCB Hookup Hint
Hi
Have a look at how close the PCB is to the output devices of a KSA50 and what short hookup wires was used compared to a KSA100 or 80B.
My gut feeling to all the potential clone builders is that
the shorter the hookup wire between the PCB and output devices the better the amp is going to sound.
Jozua
Hi
Have a look at how close the PCB is to the output devices of a KSA50 and what short hookup wires was used compared to a KSA100 or 80B.
My gut feeling to all the potential clone builders is that
the shorter the hookup wire between the PCB and output devices the better the amp is going to sound.
Jozua
Attachments
Caps on the 100Mk2 are identical to the 100Mk1 and 50Mk2. They're Roederstein 1000uF 16V 85degrees, in a hard plastic casing. The same pedigree has been used in the relay protection.
Differences between 100Mk2 and 50Mk2 are numerous. I've mentioned it on the KSA50 thread but here's a summary. It uses the same input stage, zener regulation, offset & bias adjustment, protection, feedback, wiring etc., but with a single driver stage, no cascaded pre-drivers, no Mosfets and no current diodes. Drivers are still MJE15030/MJE15031, outputs the same 15003/15004. Diff amp also 970/2240, predrivers and the other pair 968/2238. Only difference on the power stage is the emitter resistors that has been reduced to 0.5R instead of the 100Mk2's 1R. On the power supply it uses the same bridges and 75V 40,000uF Sprague caps as the 100Mk2, but with much smaller transformers that cannot be more than 500VA at the very most, probably less. Loaded rail voltages are in the area of 47VDC.
Differences between 100Mk2 and 50Mk2 are numerous. I've mentioned it on the KSA50 thread but here's a summary. It uses the same input stage, zener regulation, offset & bias adjustment, protection, feedback, wiring etc., but with a single driver stage, no cascaded pre-drivers, no Mosfets and no current diodes. Drivers are still MJE15030/MJE15031, outputs the same 15003/15004. Diff amp also 970/2240, predrivers and the other pair 968/2238. Only difference on the power stage is the emitter resistors that has been reduced to 0.5R instead of the 100Mk2's 1R. On the power supply it uses the same bridges and 75V 40,000uF Sprague caps as the 100Mk2, but with much smaller transformers that cannot be more than 500VA at the very most, probably less. Loaded rail voltages are in the area of 47VDC.
