Thanks for those. I know how to insulate the TO3 device from the heatsink. My concern is that the collector on the bolt (on the +ive supply), from the top of the device is connected to a metal standoff that mounts the main PCB on a pad that is connected to the +ive supply. That gives two power collections between the two boards. And that the standoff is "live".
I'm using this:
Security Check
Insulating bushing is shown at the bottom of the page.
Mica washer with two bushings cost just 0.14€. I've paid €7.- for 50 pieces.
First thanks everyone here for this great discussion. I have read it several times over the past couple of months and was able to build KSA-50 from reading it only, without a single question. All conceivable questions were answered here.
I have used +30-0--30 transformer and my rails are slightly above +-42V. Currently I am testing everything at low bias. It is very likely that I will have to use forced convection to attain higher bias level because I don't want to trade-of quality for the convenience of passive cooling.
@chiili:
Today, I have successfully completed my KSA-50, without smoke. I hope you have meanwhile also solved the problem with isolating stand-offs. I am using an identical board, and if you haven't solved the problem yet, let me tell you how I did that:
I use two 3.5kg (0,3C/W) flat back heat sinks and two pairs of 15024 and 15025. I did not use nylon washers on the bolt because it must conduct to the appropriate connection on the copper side of the board. However, I have put TO-3 nylon washers into boreholes on the heat sink, to prevent contact of the board with the surface of the heat sink. Washers also prevent contact between the bolts and heat sink.
I have used +30-0--30 transformer and my rails are slightly above +-42V. Currently I am testing everything at low bias. It is very likely that I will have to use forced convection to attain higher bias level because I don't want to trade-of quality for the convenience of passive cooling.
@chiili:
Today, I have successfully completed my KSA-50, without smoke. I hope you have meanwhile also solved the problem with isolating stand-offs. I am using an identical board, and if you haven't solved the problem yet, let me tell you how I did that:
I use two 3.5kg (0,3C/W) flat back heat sinks and two pairs of 15024 and 15025. I did not use nylon washers on the bolt because it must conduct to the appropriate connection on the copper side of the board. However, I have put TO-3 nylon washers into boreholes on the heat sink, to prevent contact of the board with the surface of the heat sink. Washers also prevent contact between the bolts and heat sink.
Finally found the time to lash up one channel. Bias sets fine, offset at 0v, just letting it sit.
IMG_20200321_191522 by Garf Arf, on Flickr
IMG_20200321_191522 by Garf Arf, on Flickr
I just noticed this piece of information in your post. Please bear in mind the maximun temperature Nylon can be used at. The maximum service temperature can be as low as 80°C . You may want to look into it (if you have not already done so) .
Nice work
Cheers,
Eric
Thanks
I do indeed appreciate preparedness to help at this forum in general and a great attitude of participants in this thread in particular.
I have decided to remain at the safe side regarding temperature: 60 deg C on the heat sink with:
(1) Honeywell commercial thermostat 2455R close to transistors on the both channels.
(2) 0.3 C/W heat sink
(3) a silent fan
Calculations have indicated that it is virtually impossible to attain satisfactory biassing of this amplifier without the use of forced convection. Running this amplifier at low biass would have been an inaceptable trade-off. Hence, I am experimenting right now with suitable layout of chassis to provide the best cooling at minimum fan velocity/noise. As a model I am using a plywood chassis model which is easy to modify and adjust. When I find an adequate layout I plan to make a chassis from 3mm aluminium sheets.
I want this amplifier feel comfortable at my home.
Thanks for the warning Eric,I just noticed this piece of information in your post. Please bear in mind the maximun temperature Nylon can be used at. The maximum service temperature can be as low as 80°C . You may want to look into it (if you have not already done so) .
Nice work
Cheers,
Eric
I do indeed appreciate preparedness to help at this forum in general and a great attitude of participants in this thread in particular.
I have decided to remain at the safe side regarding temperature: 60 deg C on the heat sink with:
(1) Honeywell commercial thermostat 2455R close to transistors on the both channels.
(2) 0.3 C/W heat sink
(3) a silent fan
Calculations have indicated that it is virtually impossible to attain satisfactory biassing of this amplifier without the use of forced convection. Running this amplifier at low biass would have been an inaceptable trade-off. Hence, I am experimenting right now with suitable layout of chassis to provide the best cooling at minimum fan velocity/noise. As a model I am using a plywood chassis model which is easy to modify and adjust. When I find an adequate layout I plan to make a chassis from 3mm aluminium sheets.
I want this amplifier feel comfortable at my home.
This quote is from another tread, but this belongs here
Your quoted post is the best description of the problem of biassing this amplifier whilst maintaining the need for forced convection at minimum.
Initially I hoped that passive cooling would have been just enough but it has put too stringent constraint on biassing current.
Running this fine amplifier at low biass seems to me like buying a fine stallion and castrating him out of hand. It sounds the best hot, unleashed.
I have managed to keep temperatures under control with minimum aid of forced convection whilst maintaining bias current at 450 mA per device. To achieve this I have used airtight chassis and have left just three openings:
(1) on the back of the chassis there is an opening for a 120mm fan, for the entrance of air
(2) two narrow slots below heat sinks as exaust
My objextive was not to create excessive flow rates, rather to create well controlled air flow throughout the entire chassis.
Photo 1 is presents my initial design concept - a simple steel skeleton. In front you may see my MIG welding gun and at tle left there is my welding mask. The rough surface under the skeleton is my steel welding bench. This steel skeleton is already scrapped.
Now, imagine these two sinks shifted towards the center of the skeleton. With this confoguration all components are well exposed to the air flow which enters horizontally.
I have used »be quiet« Shadow Wings 2 120mm fan operating at reduced voltage, at 10VDC. It is dead silent at reduced voltage. There are two slots below the heat sinks. I have fixed the both driver boards on top of the chassis with with components facing the bottom of the chassis.On the photo 2 you may see components in a plywood model, turned upside-down. During operation the chassis is sealed.
Output stage is fixed on two flat back heat sinks with dimensions 300x200x40 mm. The fan blows air across all components and air is well mixed. The point of this approach is to create enough turbulences inside the chassis to make the film of hot air at surfaces of components generating heat as thin as possible. Above the heat sinks I have left 1/4 inch slot to direct the airflow towards the outer side where narrow slots are located to direct the air out of the chassis in a controlled way across the heat sinks. Please note, that it is not necessary to blow the air directly towards the heat sinks. That is in fact counterproductive because direct blowing of air against the heat sinks locally creates channels of high air velocity while leaving dead-ends on other parts of heat sink, with zero flow rate. What really matters is to create an overpressure within the chamber and then let the air spontaneously flow through a thin slot. This creates an uniform plug-flow across the entire heat exchange area. In turn, this reduces thickness of hot air otherwise blocking heat exchange. In this way it is possible to displace all air from the chassis in an optimal way whilst maintaining increased heat transfer.
I have tested this on a plywood model and the result is well tuned amplifier. Now I just wait shops to be re-opened after this Corona maddness to purchase aluminium sheets and make a decent chasis for this amplifier.
Cheers
Alan,
Your quoted post is the best description of the problem of biassing this amplifier whilst maintaining the need for forced convection at minimum.
Initially I hoped that passive cooling would have been just enough but it has put too stringent constraint on biassing current.
Running this fine amplifier at low biass seems to me like buying a fine stallion and castrating him out of hand. It sounds the best hot, unleashed.
I have managed to keep temperatures under control with minimum aid of forced convection whilst maintaining bias current at 450 mA per device. To achieve this I have used airtight chassis and have left just three openings:
(1) on the back of the chassis there is an opening for a 120mm fan, for the entrance of air
(2) two narrow slots below heat sinks as exaust
My objextive was not to create excessive flow rates, rather to create well controlled air flow throughout the entire chassis.
Photo 1 is presents my initial design concept - a simple steel skeleton. In front you may see my MIG welding gun and at tle left there is my welding mask. The rough surface under the skeleton is my steel welding bench. This steel skeleton is already scrapped.
Now, imagine these two sinks shifted towards the center of the skeleton. With this confoguration all components are well exposed to the air flow which enters horizontally.
I have used »be quiet« Shadow Wings 2 120mm fan operating at reduced voltage, at 10VDC. It is dead silent at reduced voltage. There are two slots below the heat sinks. I have fixed the both driver boards on top of the chassis with with components facing the bottom of the chassis.On the photo 2 you may see components in a plywood model, turned upside-down. During operation the chassis is sealed.
Output stage is fixed on two flat back heat sinks with dimensions 300x200x40 mm. The fan blows air across all components and air is well mixed. The point of this approach is to create enough turbulences inside the chassis to make the film of hot air at surfaces of components generating heat as thin as possible. Above the heat sinks I have left 1/4 inch slot to direct the airflow towards the outer side where narrow slots are located to direct the air out of the chassis in a controlled way across the heat sinks. Please note, that it is not necessary to blow the air directly towards the heat sinks. That is in fact counterproductive because direct blowing of air against the heat sinks locally creates channels of high air velocity while leaving dead-ends on other parts of heat sink, with zero flow rate. What really matters is to create an overpressure within the chamber and then let the air spontaneously flow through a thin slot. This creates an uniform plug-flow across the entire heat exchange area. In turn, this reduces thickness of hot air otherwise blocking heat exchange. In this way it is possible to displace all air from the chassis in an optimal way whilst maintaining increased heat transfer.
I have tested this on a plywood model and the result is well tuned amplifier. Now I just wait shops to be re-opened after this Corona maddness to purchase aluminium sheets and make a decent chasis for this amplifier.
Cheers
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Building Guide
Please note that mje15032 is TO−220 device. I have used MJ15024/15025. There are multiple alternatives. For more details look into: Krell KSA-50 clone Building Guide
Last edited:
thanks for the answer, I looked at krell guide, but specifically I did not find an alternative to 2sc3955 of q111
Ah I understand now,
we were looking at two different schematiscs. I yours q111 is MJE15030. Take into account that 2SC3955 is/was TO-126ML and MJE15030 is TO-220.
Last edited:
Ah I understand now,
I have mje 15032 is teorical correct sostitute, and base resistor , 2,2 ohm is necessary ?
There are four things that could be usefull for you:
1. Regarding Q111 substitute you are correct about MJE15032
2. Partlist by pinkmouse
3. Original schematics (in the middle of the schematics there are individual devices.
4. Revised schematic published at this forum, I forgot where but I have download it. It is attached below.
Regarding the 2.2 resistor I am not sure because there is no 2.2 resistor in any of part lists of KSA-50 I have ever seen.
I am sure there are other experienced experts arround here who should help you to complete this project. My present day knowledge isn't sufficient to do that adequately, but in few months I will improve - I'm a quick learner.
I wish you lots of luck and health which you will need these days in Italy.
In bocca al lupo!
(Eng. Good luck)
Last edited:
There are four things that could be usefull for you:
1. Regarding Q111 substitute you are correct about MJE15032
2. Partlist by pinkmouse
3. Original schematics (in the middle of the schematics there are individual devices.
4. Revised schematic published at this forum, I forgot where but I have download it. It is attached below.
Regarding the 2.2 resistor I am not sure because there is no 2.2 resistor in any of part lists of KSA-50 I have ever seen.
I am sure there are other experienced experts arround here who should help you to complete this project. My present day knowledge isn't sufficient to do that adequately, but in few months I will improve - I'm a quick learner.
I wish you lots of luck and health which you will need these days in Italy.
In bocca al lupo!
Thank’ s for your information
the situation in Italy is difficult, but let's move on, thanks for the support.
One more thing regarding that 2.2 resistor. I have seen it now at the original Delta-Audio scheme but recent schematics have addopted 0R5 or 0R68. You will deal with these resistors while adjusting biass.
Cheers!
