@Chris Hornbeck
Many thanks for this very helpful explanation.
My deep and extensive reading on this subject has not shed this sort of illumination.
Appreciated.
Many thanks for this very helpful explanation.
My deep and extensive reading on this subject has not shed this sort of illumination.
Appreciated.
It has been several months since I invested in a matched quartet of these Chinese-made SHUGUANG KT88-98 tubes.
I promised to report on them and frankly, they were pretty cheap, so I wasn't expecting too much.
Boy, was I in for a surprise!
These are excellent tubes. Extended in both bass and treble, detailed and still retaining a bit of the 6550 mid-range warmth - if perhaps not quite as lush.
What I have discovered is that each different tube I try in our beloved BEWITCH hot-rod, requires some tweaking to bring out the best in them.
That is to say, each different tube type has its own favourite operating point. It seems no two are the same.
After more than 25 different experiments, using 6550 and KT88 tubes and loads of varying operating points - I have a few dozen spreadsheets, where I have recorded every iteration and every operating point - I have come to the conclusion that there is no such thing as good and bad sounding tubes.
As long as they are healthy tubes, I now see that the circuit can be tweaked to accommodate any suitable tube.
For these KT88-98, we now run the BEWITCH with reduced Gain - using 4 x 6SN7 preamp tubes (no high-gain 6SL7 tubes in circuit, anymore) - offset by much-reduced Global Negative Feedback.
(Nod and thanks to @TonyTecson...)
We now run (small and stubby) Russian NOS 5U4C rectifier tubes which poke out considerably more power than the original Chinese 5Z4P tubes - and it suits these KT88-98s.
The (shared) cathode bias resistor remains at 250R and we only run in triode. I assume that we are getting 15W-18W per channel - which is more than enough around our way.
After much trial and error, I think we will be staying with this configuration, moving forward.
I promised to report on them and frankly, they were pretty cheap, so I wasn't expecting too much.
Boy, was I in for a surprise!
These are excellent tubes. Extended in both bass and treble, detailed and still retaining a bit of the 6550 mid-range warmth - if perhaps not quite as lush.
What I have discovered is that each different tube I try in our beloved BEWITCH hot-rod, requires some tweaking to bring out the best in them.
That is to say, each different tube type has its own favourite operating point. It seems no two are the same.
After more than 25 different experiments, using 6550 and KT88 tubes and loads of varying operating points - I have a few dozen spreadsheets, where I have recorded every iteration and every operating point - I have come to the conclusion that there is no such thing as good and bad sounding tubes.
As long as they are healthy tubes, I now see that the circuit can be tweaked to accommodate any suitable tube.
For these KT88-98, we now run the BEWITCH with reduced Gain - using 4 x 6SN7 preamp tubes (no high-gain 6SL7 tubes in circuit, anymore) - offset by much-reduced Global Negative Feedback.
(Nod and thanks to @TonyTecson...)
We now run (small and stubby) Russian NOS 5U4C rectifier tubes which poke out considerably more power than the original Chinese 5Z4P tubes - and it suits these KT88-98s.
The (shared) cathode bias resistor remains at 250R and we only run in triode. I assume that we are getting 15W-18W per channel - which is more than enough around our way.
After much trial and error, I think we will be staying with this configuration, moving forward.
Last edited:
Yup. I bought the same tubes - the cheapest quad I could find.
Frankly, I don't even remember if it was eBay or Aliexpress where I found them, but at the tube a quad was 180$CAD shipped.
I've been using them in an amp designed around the supplemental datasheel for the GEC KT88.
Cathode biased 60W amp that idles the tubes at ~40W each.
They are what they say they are: KT88 tubes. But unlike most cheap tubes, these ones last more than 6 months.
Now if only I could find a buyer for the amp they are in...
Frankly, I don't even remember if it was eBay or Aliexpress where I found them, but at the tube a quad was 180$CAD shipped.
I've been using them in an amp designed around the supplemental datasheel for the GEC KT88.
Cathode biased 60W amp that idles the tubes at ~40W each.
They are what they say they are: KT88 tubes. But unlike most cheap tubes, these ones last more than 6 months.
Now if only I could find a buyer for the amp they are in...
Hi @TonyTecson,
It has been almost a year since you and others kindly helped me find the best operating parameters of this BEWITCH 6550 power amp - currently running KT88 output tubes, and locked in TRIODE OPERATION.
[I trust you and yours are keeping well...]
As per your suggestion above, we have been very happily running the BEWITCH with 6SN7 tubes in both preamp positions, and with GNFB reduced from 100R to 240K - not sure what that equates to in terms of GNFB dB - with the GNFB resistor bypassed with 9pF silver mica cap.
It all sounds wonderful and our 93dB 3-way loudspeakers - 8"- woofers - have never sounded better.
But still, my curiosity gets the better of me!
I was thinking to try ULTRALINEAR operation again, with our new set-up.
Before trying this, I'd like to gain an expert opinion as to whether switching the amp back to UL operation could potentially damage anything?
Here is the TRIODE LOCKED circuit, as it presently operates daily and never puts a foot wrong.
Last edited:
In my (not so 🙂) expert opinion, there is not a lot of potential damage by switching between UL and triode strapped, assuming the switch happens when the amplifier is off:
- The KT88s in that schematic are biased very conservative (assuming Vak and Ia are correct). Ig2 will be a bit higher in UL during peaks, but not that much.
- Global negative feedback level is low, you will probably see bit of reduction in the damping factor for UL.
Many thanks @jcalvarez
I had come to a similar conclusion, but wanted to be sure that I did not miss anything.
Obviously, sound quality will determine if UL is worth exploring further.
Let's see...
I had come to a similar conclusion, but wanted to be sure that I did not miss anything.
Obviously, sound quality will determine if UL is worth exploring further.
Let's see...
Hello SONDEKNZ
I have to praise you for the amount of time you have invested so far and for being so keen to experiment with trying out different circuit variations and always listening to which one sounds better for you.
I have also spent the last few years improving my Bewitch 6550, both externally and internally, and also changing the circuit diagram.
I have stuck with a circuit variant based on Lampizator and will stick with it, and will not continue to look for or configure other 6550/KT88 solutions.
With our beloved BEWITCH 6550/KT88 push-pull integrated amplifier, I was always dissatisfied with how the schematic was configured and released from the factory. Recently I took the time and effort to sit in front of the PC again and looked for improvements to the circuit diagram, making changes to how the circuit board really is and changing them, and this is what the circuit looks like.
There are still a few errors on your circuit diagram and there is still room for improvement. There are still some resistors there, namely R17-R18-R19 and R20 with 100K in the power supply area, which don't actually exist on the circuit board.
Also at another point on the transformer secondary 6.3v/2A this output leads not only for the
preamp tubes
V1-V2 (6N8P) tube heater, but also the two connected resistors R23 -R24 with 100 Ohm with point A noted that the two tubes on this heating winding are increased to 42 VDC, which has a common bias voltage of only two 6550 tubes of one channel
I also find that with several capacitors that show incorrect values, instead of the 33uF/400v, a value of 22uF/400v is given on the schematic.
Please check and change on the circuit diagram if possible, unless at the places where I noted with X or circled on the circuit diagram, the components are there or you soldered them in later, then it is something different.
At least if we show schematics here for other forum users, it is also desirable and beneficial if they can find a correct schematic.
I have to praise you for the amount of time you have invested so far and for being so keen to experiment with trying out different circuit variations and always listening to which one sounds better for you.
I have also spent the last few years improving my Bewitch 6550, both externally and internally, and also changing the circuit diagram.
I have stuck with a circuit variant based on Lampizator and will stick with it, and will not continue to look for or configure other 6550/KT88 solutions.
With our beloved BEWITCH 6550/KT88 push-pull integrated amplifier, I was always dissatisfied with how the schematic was configured and released from the factory. Recently I took the time and effort to sit in front of the PC again and looked for improvements to the circuit diagram, making changes to how the circuit board really is and changing them, and this is what the circuit looks like.
There are still a few errors on your circuit diagram and there is still room for improvement. There are still some resistors there, namely R17-R18-R19 and R20 with 100K in the power supply area, which don't actually exist on the circuit board.
Also at another point on the transformer secondary 6.3v/2A this output leads not only for the
preamp tubes
V1-V2 (6N8P) tube heater, but also the two connected resistors R23 -R24 with 100 Ohm with point A noted that the two tubes on this heating winding are increased to 42 VDC, which has a common bias voltage of only two 6550 tubes of one channel
I also find that with several capacitors that show incorrect values, instead of the 33uF/400v, a value of 22uF/400v is given on the schematic.
Please check and change on the circuit diagram if possible, unless at the places where I noted with X or circled on the circuit diagram, the components are there or you soldered them in later, then it is something different.
At least if we show schematics here for other forum users, it is also desirable and beneficial if they can find a correct schematic.
"worth pursuing" ?
Differences in the "sound" of UL and Triode Wired modes ? ? ? . . . That depends, as per described below:
1. Without connecting the global negative feedback:
UL has More maximum power; Higher distortion at say 20% to 40% of its maximum power; and has Low to Medium damping factor.
Triode Wired mode has Less maximum power; Lower distortion at say 20% to 40% of maximum power; and has the Highest damping factor.
2. Disconnecting the global negative feedback:
At medium power levels you will Probably hear a difference, between UL and Triode wired modes.
3. Re-connecting the global negative feedback:
At medium power levels you will probably Not hear a difference, between UL and Triode wired modes.
With the negative feedback connected:
The maximum power levels of each operating modes do Not change.
But, the distortion is Reduced, and the damping factors are Increased.
4. I think I Will be able to hear the difference between UL and Triode wired mode if global negative feedback is Not connected.
However: I think I will Not be able to hear the difference between UL and Triode wired mode if global negative feedback Is connected.
You get the idea that I am trying to communIcate to some people.
Happy careful un-powered mode switching!
$0.03
The above accounts for inflation.
Differences in the "sound" of UL and Triode Wired modes ? ? ? . . . That depends, as per described below:
1. Without connecting the global negative feedback:
UL has More maximum power; Higher distortion at say 20% to 40% of its maximum power; and has Low to Medium damping factor.
Triode Wired mode has Less maximum power; Lower distortion at say 20% to 40% of maximum power; and has the Highest damping factor.
2. Disconnecting the global negative feedback:
At medium power levels you will Probably hear a difference, between UL and Triode wired modes.
3. Re-connecting the global negative feedback:
At medium power levels you will probably Not hear a difference, between UL and Triode wired modes.
With the negative feedback connected:
The maximum power levels of each operating modes do Not change.
But, the distortion is Reduced, and the damping factors are Increased.
4. I think I Will be able to hear the difference between UL and Triode wired mode if global negative feedback is Not connected.
However: I think I will Not be able to hear the difference between UL and Triode wired mode if global negative feedback Is connected.
You get the idea that I am trying to communIcate to some people.
Happy careful un-powered mode switching!
$0.03
The above accounts for inflation.
Last edited:
Hey @saltonm73
Welcome back to this aging thread.
Yours is a bravura response, to be sure.... Thank you.
I will study your post carefully and make the changes that apply - and I will update, below.
Many thanks for this effort. Appreciated.
-SONDEKNZ
Hey @6A3sUMMER
And welcome back to this aging thread, to you too.
As with all of your many posts on this thread, these thoughts of yours are most welcome. Many thanks for kindly sharing your deep experience with how circuit changes impact amp sound. Appreciated.
In my earlier post, I contemplated reverting to UL operation, but in the interim, returned to the original TRIODE-strapped tube configuration: 6SL7 in the preamp position 1 and 6SN7 in preamp position 2.
After running all 6SN7 tubes for so long, I immediately noticed that returning the 6SL7 tubes to the circuit improved the boogie factor - despite the fact that I continue to operate the amp with 240K (Less than factory) Global Negative Feedback. The additional Gain provided by the 6SL7 tubes is definitely beneficial to good sound.
Enjoying this return - and considering the thoughts you have shared above, I have decided NOT to switch to UL operation. Thank you for saving me much time and effort.
The old girl is sounding better than ever!
-SONDEKNZ
Last edited:
In terms of next steps, I will revisit the amp schematic - cognizant of all the good information provided by @saltonm73 - and carefully make the changes, in the hopes to correct any errors.
I have no intention in making changes to the output transformers in this amp - which was suggested much earlier in this thread - so I feel that the last bastion of exploration for me involves the main power supply capacitors and any potential variation thereof.
As the main (largest, EL) PSU caps are completely factory originals, it is in my mind that it is high time that I refresh them.
My choices include:
A: Replace like-for-like with an ultra-low ESR, highly respected audio brand; but keeping with same VDC / uF as is currently installed. (Actual values TBC once I dismantle the amp again); OR
B: Replace with an ultra-low ESR, highly respected audio brand; but increasing either (or both) VDC / uF of these EL caps.
I am aware that many more experienced experts caution against arbitrarily increasing capacitance values on these PSU caps - whilst others seem to consider it highly beneficial to the sound - especially to the bass.
I suspect each amp circuit would respond differently to these types of changes, so I would therefore appreciate any expert advice on this aspect of potential improvements to the BEWITCH 6550 (KT88) amp.
I have no intention in making changes to the output transformers in this amp - which was suggested much earlier in this thread - so I feel that the last bastion of exploration for me involves the main power supply capacitors and any potential variation thereof.
As the main (largest, EL) PSU caps are completely factory originals, it is in my mind that it is high time that I refresh them.
My choices include:
A: Replace like-for-like with an ultra-low ESR, highly respected audio brand; but keeping with same VDC / uF as is currently installed. (Actual values TBC once I dismantle the amp again); OR
B: Replace with an ultra-low ESR, highly respected audio brand; but increasing either (or both) VDC / uF of these EL caps.
I am aware that many more experienced experts caution against arbitrarily increasing capacitance values on these PSU caps - whilst others seem to consider it highly beneficial to the sound - especially to the bass.
I suspect each amp circuit would respond differently to these types of changes, so I would therefore appreciate any expert advice on this aspect of potential improvements to the BEWITCH 6550 (KT88) amp.
Last edited:
Resistors R17 to R20 were probably planned in the manufacturer's plan but the developer came to a different decision to use them in this design.
As far as I can see or as you can see online, they are not visible on any Bewitch 6550/KT88 I know of or were not even soldered to the circuit board, and probably were not absolutely necessary to use them either because
current can also discharge via the other existing resistors.
That's why I asked SONDEKNZ whether he had perhaps subsequently soldered them onto the circuit board!!
Saltonm73
As far as I can see or as you can see online, they are not visible on any Bewitch 6550/KT88 I know of or were not even soldered to the circuit board, and probably were not absolutely necessary to use them either because
current can also discharge via the other existing resistors.
That's why I asked SONDEKNZ whether he had perhaps subsequently soldered them onto the circuit board!!
Saltonm73
@saltonm73
@MarekH
Thanks for the input.
Below is a photo of the main BEWITCH 6550 PCB - not mine! - showing two PSU 100K resistors that are definitely in operation.
Whether they are R17, R18, R19 or R20, I have not determined.
So, it is in my thinking that two of the four PSU resistors shown on the factory schematic are missing - not all four.
@MarekH
Thanks for the input.
Below is a photo of the main BEWITCH 6550 PCB - not mine! - showing two PSU 100K resistors that are definitely in operation.
Whether they are R17, R18, R19 or R20, I have not determined.
So, it is in my thinking that two of the four PSU resistors shown on the factory schematic are missing - not all four.
hallo SONDEKNZ
Die eingekreisten Stellen, die Widerstände R23 und R24 sind 100 Ohm, sind sie nicht, ich habe das hier schon einmal erwähnt, gehen Sie ein paar Seiten zurück https://www.diyaudio.com/community/threads/bewitch-kt88.226866 /post-6909748 aber scheinbar schaut sich niemand meinen Schaltplan und die Bilder richtig an. Bei R23 und R24 liegt der Transformator sekundär 6,3V/2A. Dieser Ausgang führt nicht nur zu den Vorstufenröhren V1-V2 (6N8P) Röhrenheizung, sondern auch zu den beiden angeschlossenen Widerständen R23 -R24 mit 100 Ohm bei Punkt A. Beachten Sie, dass die beiden Röhren 6550 an dieser Heizwicklung auf 42 VDC erhöht werden, was eine gemeinsame Vorspannung von nur zwei 6550 Röhren eines linken Kanals hat, für den rechten Kanal ist nichts vorgesehen.
Die eingekreisten Stellen, die Widerstände R23 und R24 sind 100 Ohm, sind sie nicht, ich habe das hier schon einmal erwähnt, gehen Sie ein paar Seiten zurück https://www.diyaudio.com/community/threads/bewitch-kt88.226866 /post-6909748 aber scheinbar schaut sich niemand meinen Schaltplan und die Bilder richtig an. Bei R23 und R24 liegt der Transformator sekundär 6,3V/2A. Dieser Ausgang führt nicht nur zu den Vorstufenröhren V1-V2 (6N8P) Röhrenheizung, sondern auch zu den beiden angeschlossenen Widerständen R23 -R24 mit 100 Ohm bei Punkt A. Beachten Sie, dass die beiden Röhren 6550 an dieser Heizwicklung auf 42 VDC erhöht werden, was eine gemeinsame Vorspannung von nur zwei 6550 Röhren eines linken Kanals hat, für den rechten Kanal ist nichts vorgesehen.
Last edited:
Hallo nochmal SONDEKNZ
Ich habe mir mal dein altes Bild aufgerufen und angeschaut, welche Widerstände (R23-R24) auf deiner Platine verbaut bzw. ausgetauscht wurden, der Wert ist mit 100 Ohm korrekt, es handelt sich also nicht um 100K Widerstände wie oben erwähnt.
https://www.diyaudio.com/community/threads/bewitch-kt88.226866/post-6594478
dann sieht man auch das es nicht die R17 - R20 ist.
Ich habe mir mal dein altes Bild aufgerufen und angeschaut, welche Widerstände (R23-R24) auf deiner Platine verbaut bzw. ausgetauscht wurden, der Wert ist mit 100 Ohm korrekt, es handelt sich also nicht um 100K Widerstände wie oben erwähnt.
https://www.diyaudio.com/community/threads/bewitch-kt88.226866/post-6594478
dann sieht man auch das es nicht die R17 - R20 ist.
Hello,
I just noticed that when I copied my text, it was written in German by mistake, please excuse me
here again in English
" hello SONDEKNZ
The circled places, the resistors R23 and R24 are 100 ohms, they are not, I have already mentioned this here in the past, go back a few pages https://www.diyaudio.com/community/threads/bewitch-kt88.226866/post-6909748
but apparently
no one is even looking properly at my circuit diagram and pictures.
The R23 and R24 are where the transformer secondary 6.3v/2A this output leads not only for the
preamp tubes
V1-V2 (6N8P) tube heater, but also to the two connected resistors R23 -R24 with 100 Ohm with point A noted that the two tubes 6550 on this heating winding are increased to 42 VDC which has a common bias voltage of only two 6550 tubes of a left channel, nothing is provided for the right channel.
" Hello again SONDEKNZ
I accessed your old picture and looked at which resistors (R23-R24) were used or replaced on your board, the value is correct at 100 Ohm, so they are not 100K resistors as mentioned above, then you can also see that it is not the R17 - R20.
I just noticed that when I copied my text, it was written in German by mistake, please excuse me
here again in English
" hello SONDEKNZ
The circled places, the resistors R23 and R24 are 100 ohms, they are not, I have already mentioned this here in the past, go back a few pages https://www.diyaudio.com/community/threads/bewitch-kt88.226866/post-6909748
but apparently
no one is even looking properly at my circuit diagram and pictures.
The R23 and R24 are where the transformer secondary 6.3v/2A this output leads not only for the
preamp tubes
V1-V2 (6N8P) tube heater, but also to the two connected resistors R23 -R24 with 100 Ohm with point A noted that the two tubes 6550 on this heating winding are increased to 42 VDC which has a common bias voltage of only two 6550 tubes of a left channel, nothing is provided for the right channel.
" Hello again SONDEKNZ
I accessed your old picture and looked at which resistors (R23-R24) were used or replaced on your board, the value is correct at 100 Ohm, so they are not 100K resistors as mentioned above, then you can also see that it is not the R17 - R20.