Pierre,
I think the difference you'll find in the 50 Mk-2 is the addition of the differential FETS like in the 100 Mk-2. If this is really the case it should be entirely possible to build the 50Mk-2 on the current 100 MK-2 PCB just by leaving out the extra pair of drivers and changing some component values. Looking forword to seeing the drawing!!
Mark
I think the difference you'll find in the 50 Mk-2 is the addition of the differential FETS like in the 100 Mk-2. If this is really the case it should be entirely possible to build the 50Mk-2 on the current 100 MK-2 PCB just by leaving out the extra pair of drivers and changing some component values. Looking forword to seeing the drawing!!
Mark
Entirely possible for the bridging.. the thing on the outside has a slot so perhaps it's a switch that has to be turned with a screwdriver. Neat little trick but I wonder how the noise would be..
Sorry Mark, no FET's on this beast - and upon request from Krell its serial number was confirmed as a Mk2. The only semiconductors per board are:
2x 2SC2240
2x 2SA970
2x 2SC2238
2x 2SA968
2x MJE15030
1x MJE15031
2x zeners (not sure of the voltage yet)
2x diodes (probably 1N914 or similar)
Capacitors from what I could see was just the normal input RF cap, Miller caps, 100nF slowdown cap, feedback caps and then the weird 2.2uF. Resistors are normal Dale, and all I could get to with the meter were values corresponding to the clone's. Close to the input there are two resistors per board tack-soldered, one in a pretty little sleeve.
Gauging from that it would probably be easier to mod a KSA50 clone board than a 100.2 one. The rather high DC rails was a bit of a surprise, AFAIK most people went for mid-30's.
Sorry Mark, no FET's on this beast - and upon request from Krell its serial number was confirmed as a Mk2. The only semiconductors per board are:
2x 2SC2240
2x 2SA970
2x 2SC2238
2x 2SA968
2x MJE15030
1x MJE15031
2x zeners (not sure of the voltage yet)
2x diodes (probably 1N914 or similar)
Capacitors from what I could see was just the normal input RF cap, Miller caps, 100nF slowdown cap, feedback caps and then the weird 2.2uF. Resistors are normal Dale, and all I could get to with the meter were values corresponding to the clone's. Close to the input there are two resistors per board tack-soldered, one in a pretty little sleeve.
Gauging from that it would probably be easier to mod a KSA50 clone board than a 100.2 one. The rather high DC rails was a bit of a surprise, AFAIK most people went for mid-30's.
Exploring a KSA 50.2
Hi
I am intrigued by the sonic differences between my clone 50 and the 20 year old KSA50.2 At this stage the KSA50.2 runs rings around the clone i.t.o. soundstage imagery, bass deepness and overall sweetness.
The only major difference between the two amps is that my clone uses a lower rail voltage and much higher quality parts on the pcb versus the standard Dale resistors and Roederstein caps on the KSA50.2
I want to know what needs to be done to get the clone to sound exactly or even better than the standard KSA50.2? In theory the clone should be better but it is not.
Jozua
Hi
I am intrigued by the sonic differences between my clone 50 and the 20 year old KSA50.2 At this stage the KSA50.2 runs rings around the clone i.t.o. soundstage imagery, bass deepness and overall sweetness.
The only major difference between the two amps is that my clone uses a lower rail voltage and much higher quality parts on the pcb versus the standard Dale resistors and Roederstein caps on the KSA50.2
I want to know what needs to be done to get the clone to sound exactly or even better than the standard KSA50.2? In theory the clone should be better but it is not.
Jozua
Mark,
Nope the four devices on the lower heatsink are all 2SC2238 and 2SA968. If the circuit diagram is the same as the clone (which so far seems likely), then the extra pair is used instead of the clone's MPSA42/92, and can the clone be adapted with a bit of ingenious soldering and adding the extra bobs&bits. Even if there were some other changes to the diagram there are enough similarities to convert a clone without molesting the board too much.
Raising the rails would also help perhaps, there must be a reason why Krell had gone so high for the 50Mk2 - after all, those 75V caps are the same as used on the 100. Did anyone experiment with different voltage rails at the same bias?
Jozua, the 30V rails on your clone is almost 40% lower and sags a lot under load so in fairness the power supply isn't on par.
Nope the four devices on the lower heatsink are all 2SC2238 and 2SA968. If the circuit diagram is the same as the clone (which so far seems likely), then the extra pair is used instead of the clone's MPSA42/92, and can the clone be adapted with a bit of ingenious soldering and adding the extra bobs&bits. Even if there were some other changes to the diagram there are enough similarities to convert a clone without molesting the board too much.
Raising the rails would also help perhaps, there must be a reason why Krell had gone so high for the 50Mk2 - after all, those 75V caps are the same as used on the 100. Did anyone experiment with different voltage rails at the same bias?
Jozua, the 30V rails on your clone is almost 40% lower and sags a lot under load so in fairness the power supply isn't on par.
Nope, that's not what's going on in the Mk2:
1) Jan used 2SC2240 & 2SA970 for the input stage; same as the original and actually all the other KSA's I've ever come across.
2) Jan used MJE15030 for both driver (Q109&Q110) and pre-driver (Q107&Q108). The Mk2 use these only for driver. Now we know from the KSA100 Mk1 and Mk2 that Krell used 2SA968 and 2SC2238 for pre-drivers, so that accounts for that.
3) This is where the similarities stop. The differences from here on are that the Mk2 has an extra pair of 2SC2238 and 2SA968 (absent on the Klone), while the Klone has the MPSA42/92 (absent on the Mk2).
Since so many other variables are the same such as the 1k5 and 24R resistor values I measured, it may be possible that one should just replace the MPSA42/92 (and maybe pre-drivers too) with 2SC2238 and 2SA968, but as I said there are other sinister tack-solderings also going on that is perhaps not documented (refer to the 20pF cap in the feedback on the 100 Mk2). Furthermore I'm not sure if the zeners would still be 27V with the same resistor values with the 49V rails.
I'm not sure when, but I've to do some restoration on the boards sometime and will draw the entire new circuit diagram. Getting the actual values for e.g. C105&106 will also be nice to build the amp exactly as the DanMan intended to
1) Jan used 2SC2240 & 2SA970 for the input stage; same as the original and actually all the other KSA's I've ever come across.
2) Jan used MJE15030 for both driver (Q109&Q110) and pre-driver (Q107&Q108). The Mk2 use these only for driver. Now we know from the KSA100 Mk1 and Mk2 that Krell used 2SA968 and 2SC2238 for pre-drivers, so that accounts for that.
3) This is where the similarities stop. The differences from here on are that the Mk2 has an extra pair of 2SC2238 and 2SA968 (absent on the Klone), while the Klone has the MPSA42/92 (absent on the Mk2).
Since so many other variables are the same such as the 1k5 and 24R resistor values I measured, it may be possible that one should just replace the MPSA42/92 (and maybe pre-drivers too) with 2SC2238 and 2SA968, but as I said there are other sinister tack-solderings also going on that is perhaps not documented (refer to the 20pF cap in the feedback on the 100 Mk2). Furthermore I'm not sure if the zeners would still be 27V with the same resistor values with the 49V rails.
I'm not sure when, but I've to do some restoration on the boards sometime and will draw the entire new circuit diagram. Getting the actual values for e.g. C105&106 will also be nice to build the amp exactly as the DanMan intended to
Dear people,
Without raising a storm, just a poll about what the interest would be of a v3 Klone PCB GB. Not that there's anything wrong with PinkMouse's boards, but it appears that the transistor choice has a substantial influence on the sound and that the originals fare better.
There are also a few extra sinister components on the original that're absent on the clone, and I'll be stripping a KSA50 Mk2 soon to re-draw a 100% circuit diagram, and can implement these on the new boards. An option would be something compatible with the Pinkmouse boards for a drop-in replacement, or otherwise with the original (as I've done on the KSA100 Mk2 boards) for a more authentic look.
Mark has already done more than his part with KSA GB's, but depending on interest we can see what can be arranged with someone else handling the order.
Pierre
Without raising a storm, just a poll about what the interest would be of a v3 Klone PCB GB. Not that there's anything wrong with PinkMouse's boards, but it appears that the transistor choice has a substantial influence on the sound and that the originals fare better.
There are also a few extra sinister components on the original that're absent on the clone, and I'll be stripping a KSA50 Mk2 soon to re-draw a 100% circuit diagram, and can implement these on the new boards. An option would be something compatible with the Pinkmouse boards for a drop-in replacement, or otherwise with the original (as I've done on the KSA100 Mk2 boards) for a more authentic look.
Mark has already done more than his part with KSA GB's, but depending on interest we can see what can be arranged with someone else handling the order.
Pierre
Hi,
insulate one leg of the 2s types with a tube cut from 0.6mm insulated solid core.
Insert the three legs into the relevant holes in the pink PCB.
The mpsa types are better replaced by bc types, 550/560c do a good job. Does anyone use mpsa types? I have looked at a few different versions and there are always better performing transistors that I think do the job better.
Q105/106 replace with bc546c/556c.
insulate one leg of the 2s types with a tube cut from 0.6mm insulated solid core.
Insert the three legs into the relevant holes in the pink PCB.
The mpsa types are better replaced by bc types, 550/560c do a good job. Does anyone use mpsa types? I have looked at a few different versions and there are always better performing transistors that I think do the job better.
Q105/106 replace with bc546c/556c.
PWatts said:
A drop-in board in Watts 100MKII style, either for the authentic or the revised circuit, may push the Pink-50 boards into oblivion.
Problem with that is that Mark likely still has a bunch of Pinky boards.
Considering Mark's involvement/efforts/enthousiasm with everything related to the 50MKII/100MKII, fair would be that not a single V3 board leaves a V2 board lying in the Gulbrandsen closet.
A by-pass would be to restrict an initial gb to the ones who already have a V2.
Sure me is keen on replacing a V2 front end board by one as dandy looking as the 100MKII, without any twisting and sleeving.
(@GWB, we is to educate the children)
Attachments
For a small batch (e.g. <20) I'm willing to do the admin but unfortunately PayPal with South Africa is like the anode of diode current: funds are only allowed outward 
Adapting the KSA100 clone board layout to a KSA50 should be fairly simple unless there's an outcry for extra bells & whistles like a USB port
Support for various transistors as on the 100 without tack-soldering and hacking would also help.
BUT: whoever handles it would have to ensure that keen parties swear on their parents' graves and subsequent inheritence that it will be built and used, not stuffed in a closet for someday when there's time/money/resources/willpower.
Of all the PM and Delta boards sold, how many have actually been populated, and how many of them are in a chassis connected to a preamp and speakers?
Adapting the KSA100 clone board layout to a KSA50 should be fairly simple unless there's an outcry for extra bells & whistles like a USB port
Support for various transistors as on the 100 without tack-soldering and hacking would also help.BUT: whoever handles it would have to ensure that keen parties swear on their parents' graves and subsequent inheritence that it will be built and used, not stuffed in a closet for someday when there's time/money/resources/willpower.
Of all the PM and Delta boards sold, how many have actually been populated, and how many of them are in a chassis connected to a preamp and speakers?
