KSA 100 from Jim's Audio on EBAY - Page 4

Kaplaars · 03:29 PM

@xslavic thanks for your reply! I don't think the rectifier is too slow for the job. These are rectifiers which can handle 35A (MB3510). Almost the same as used in the original KSA-100, which works great. There is no need for faster diodes like schottky's. Schottky's were never designed for large currents, so it gets pretty expensive and difficult to get schottky's which can handle the current this amplifier draws. The only advantage for schottky's which I see is the lower voltage drop, but that's not a problem (got about 2*49V DC per channel). This is the MKII by the way :-)

@Dean: thanks for your reply! I am curious if more people used the same devices.

@Purtle: I've bought them second hand from a guy who stipped two pretty big PSU's down. There were a lot BUX81's on the cooling tunnels. That is why I did not mount the MJ15003/4's direct to the PCB's and chose this configuration; there are a few holes to much. Now they are covered :-) I have posted some pictures in this topic of how the tunnel coolers were before I striped them.

Kaplaars · 12th May 2012, 02:54 PM

Had some spare time, which does not occur that often, so I have made some further progress :-)! I've mounted the speaker protection board and the softstart I've bought earlier:

I bought the soft start via Ebay (Class A power delay soft-start temperature protection board 110V/220V | eBay). It is a nice soft start which has also two terminals for two thermostarts. The thermostats (TO-220 housing) were included. At 75 degrees Celsius the soft start cuts the power due to shorting of the thermostats. Maybe I buy thermostats which shorts at higher temperature, because 75 degrees is not that uncommon for class A amplifiers (the heatsinks of my 60W AudioAnalyse A9 reach almost that temperature and that amplifier is not even biassed full class A!).

I had some serious problems with the soft start, it is pretty crappy designed. It was not proper protected against switch debouncing, so it did not turn on or shut down right. The problem was caused by two small bad 1uF capacitors on the board, which seem to be bad quality. Changed them for Philips 1uF capacitors and the board works like it should work. But there is a but. They did a strange thing, they mounted a capacitor in series with the AC mains, which is later bypassed by a relay (together with the NTC's). Why did they do this? Is this to limit the current which flows during power on? Problem is that they mounted it in such a way that it continue bleeds about 55VAC. I changed the value of this capacitor to 22nF (X1, Y2 specified) so it bleeds less AC, but it is still about 5VAC. Is there a way to change this? I can't cut the capacitor out, because the relays than start to bounce. Sorry for the bad schematic I drew with paint, but what they did is basicly this:

At power on relay 1 switches and a few seconds later relay 2 switches.

I have also another question. Some 'high-enders' bypass the electrolytic caps with 1 - 4.7uF capacitors. They place them paralel over every electrolytic capacitor. Is this to repress HF distortion from outside? Should I place them aswell? I have some nice 1uF WIMA capacitors, so to obtain them is not a problem.

Below is a picture of how I want to place all of the components (I know it is still a mess

)

AndrewT · 12th May 2012, 04:52 PM

Do NOT fit NTC in parallel.
If you need more dissipation capacity than one NTC can manage then series connect them.

I don't like the risk you are taking by using caps across the rectifier diodes. They may be MKS which have slightly higher esr than MKP.
I think the risk of ringing and/or oscillation is too high using caps alone instead of a true snubber, i.e. R+C

Kaplaars · 12th May 2012, 05:13 PM

Hi Andrew,

The little schematic I drew was not entirely correct, what they did on the PCB was two NTC's parallel and than set two of these sets in series. So in total there are 4 NTC. My main concern is the capacitor which is mounted in series witch the mains. I do not understand what the function of this is.

What would be the risk of the snubber capacitors across the bridge rectifiers? It is a very commonly used concept. An accuphase E204 for example uses the same concept. Altough this amplifier has ceramic capacitors across the bridge rectifier. The capacitors I use are MKT and rated for 100VAC which should be enough.

AndrewT · 12th May 2012, 05:19 PM

I will post the info again:

Quote:

Do NOT fit NTC in parallel.

Mains capacitors must be mains rated. They must have either the "X" or "Y" rating on them.

Kaplaars · 12th May 2012, 05:25 PM

Hi Andrew,

Why is it a bad idea for NTC's to be mounted in parallel?

I know the mains capacitor has to be X or Y rated. Like I posted earlier the capacitor that is in series with the mains is x1, y2 rated. But that is not really my question. I dont understand why this capacitor is placed in series with the mains. I also do not entirely understand what is wrong with the C's parallel to the bridge. Can you explain that to me?

AndrewT · 12th May 2012, 05:26 PM

Quote:

Why is it a bad idea for NTC's to be mounted in series?

Read english !

Kaplaars · 12th May 2012, 05:37 PM

That is not a constructive reply Andrew :-( Made a typo, I meant parallel. Sorry, I am only human.

But you still have not answered my question(s).

CBS240 · 12th May 2012, 06:02 PM

NTC - negative temperature coefficient. When temp rises resistance lowers. If they are in parallel the resistance may drop too low. The R-T curves will no longer apply as designed. For more Pd use them in series.

Using caps alone across the rectifier (dielectric type is important for this app) creates damping with no where for the damped charge to dissipate. It needs a resistor to absorb the stray energy from the switching noise of the diodes, IOW R-C filter. Using faster diodes increases switching noise. Some of the ultra-fast or Schottky rectifiers are not so suited for 50-60Hz, more for SMPS switching speeds.

Kaplaars · 12th May 2012, 06:20 PM

Hi CBS240,

Thanks for your explanation

So actually the NTC design in the softstart is ok, because they placed two NTC's parallel and then two of this sets in series. The total resistance would than be equal to one lone NTC. For example 1 NTC is 100 ohms, two of them parallel would give 50 ohms, but if you set two of this parallel sets in series the sum is again 100 ohms. In this case setting NTC's parallel is no problem, or am I wrong?

Aha that is a good point, the absorbed energy in the capacitors has no where to go. So actually how I mounted them now is nog dangerous, but not really functional?

9th May 2012, 03:24 PM	#31
Kaplaars diyAudio Member Join Date: Jul 2010 Location: Wageningen	@xslavic thanks for your reply! I don't think the rectifier is too slow for the job. These are rectifiers which can handle 35A (MB3510). Almost the same as used in the original KSA-100, which works great. There is no need for faster diodes like schottky's. Schottky's were never designed for large currents, so it gets pretty expensive and difficult to get schottky's which can handle the current this amplifier draws. The only advantage for schottky's which I see is the lower voltage drop, but that's not a problem (got about 249V DC per channel). This is the MKII by the way :-) @Dean: thanks for your reply! I am curious if more people used the same devices. @Purtle: I've bought them second hand from a guy who stipped two pretty big PSU's down. There were a lot BUX81's on the cooling tunnels. That is why I did not mount the MJ15003/4's direct to the PCB's and chose this configuration; there are a few holes to much. Now they are covered :-) I have posted some pictures in this topic of how the tunnel coolers were before I striped them. Last edited by Kaplaars; 9th May 2012 at 03:29 PM.*

12th May 2012, 06:02 PM	#39
CBS240 diyAudio Member Join Date: Aug 2005 Location: K-town	NTC - negative temperature coefficient. When temp rises resistance lowers. If they are in parallel the resistance may drop too low. The R-T curves will no longer apply as designed. For more Pd use them in series. Using caps alone across the rectifier (dielectric type is important for this app) creates damping with no where for the damped charge to dissipate. It needs a resistor to absorb the stray energy from the switching noise of the diodes, IOW R-C filter. Using faster diodes increases switching noise. Some of the ultra-fast or Schottky rectifiers are not so suited for 50-60Hz, more for SMPS switching speeds. __________________ All the trouble I've ever been in started out as fun......

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
about ksa-100	dimis	Solid State	3	6th February 2005 07:55 PM
Help for repair my KSA 100	Alex helge	Solid State	76	23rd October 2003 03:38 PM
Bias KSA 100	Alex helge	Solid State	1	24th August 2003 11:19 AM
Krell KSA 100	Alex helge	Solid State	2	24th February 2003 03:05 AM

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

12th May 2012, 02:54 PM	#32
Kaplaars diyAudio Member Join Date: Jul 2010 Location: Wageningen	Had some spare time, which does not occur that often, so I have made some further progress :-)! I've mounted the speaker protection board and the softstart I've bought earlier: I bought the soft start via Ebay (Class A power delay soft-start temperature protection board 110V/220V \| eBay). It is a nice soft start which has also two terminals for two thermostarts. The thermostats (TO-220 housing) were included. At 75 degrees Celsius the soft start cuts the power due to shorting of the thermostats. Maybe I buy thermostats which shorts at higher temperature, because 75 degrees is not that uncommon for class A amplifiers (the heatsinks of my 60W AudioAnalyse A9 reach almost that temperature and that amplifier is not even biassed full class A!). I had some serious problems with the soft start, it is pretty crappy designed. It was not proper protected against switch debouncing, so it did not turn on or shut down right. The problem was caused by two small bad 1uF capacitors on the board, which seem to be bad quality. Changed them for Philips 1uF capacitors and the board works like it should work. But there is a but. They did a strange thing, they mounted a capacitor in series with the AC mains, which is later bypassed by a relay (together with the NTC's). Why did they do this? Is this to limit the current which flows during power on? Problem is that they mounted it in such a way that it continue bleeds about 55VAC. I changed the value of this capacitor to 22nF (X1, Y2 specified) so it bleeds less AC, but it is still about 5VAC. Is there a way to change this? I can't cut the capacitor out, because the relays than start to bounce. Sorry for the bad schematic I drew with paint, but what they did is basicly this: At power on relay 1 switches and a few seconds later relay 2 switches. I have also another question. Some 'high-enders' bypass the electrolytic caps with 1 - 4.7uF capacitors. They place them paralel over every electrolytic capacitor. Is this to repress HF distortion from outside? Should I place them aswell? I have some nice 1uF WIMA capacitors, so to obtain them is not a problem. Below is a picture of how I want to place all of the components (I know it is still a mess )

12th May 2012, 04:52 PM	#33
AndrewT diyAudio Member Join Date: Jul 2004 Location: Scottish Borders	Do NOT fit NTC in parallel. If you need more dissipation capacity than one NTC can manage then series connect them. I don't like the risk you are taking by using caps across the rectifier diodes. They may be MKS which have slightly higher esr than MKP. I think the risk of ringing and/or oscillation is too high using caps alone instead of a true snubber, i.e. R+C

12th May 2012, 05:13 PM	#34
Kaplaars diyAudio Member Join Date: Jul 2010 Location: Wageningen	Hi Andrew, The little schematic I drew was not entirely correct, what they did on the PCB was two NTC's parallel and than set two of these sets in series. So in total there are 4 NTC. My main concern is the capacitor which is mounted in series witch the mains. I do not understand what the function of this is. What would be the risk of the snubber capacitors across the bridge rectifiers? It is a very commonly used concept. An accuphase E204 for example uses the same concept. Altough this amplifier has ceramic capacitors across the bridge rectifier. The capacitors I use are MKT and rated for 100VAC which should be enough.

12th May 2012, 05:25 PM	#36
Kaplaars diyAudio Member Join Date: Jul 2010 Location: Wageningen	Hi Andrew, Why is it a bad idea for NTC's to be mounted in parallel? I know the mains capacitor has to be X or Y rated. Like I posted earlier the capacitor that is in series with the mains is x1, y2 rated. But that is not really my question. I dont understand why this capacitor is placed in series with the mains. I also do not entirely understand what is wrong with the C's parallel to the bridge. Can you explain that to me?

12th May 2012, 05:37 PM	#38
Kaplaars diyAudio Member Join Date: Jul 2010 Location: Wageningen	That is not a constructive reply Andrew :-( Made a typo, I meant parallel. Sorry, I am only human. But you still have not answered my question(s).

12th May 2012, 06:20 PM	#40
Kaplaars diyAudio Member Join Date: Jul 2010 Location: Wageningen	Hi CBS240, Thanks for your explanation So actually the NTC design in the softstart is ok, because they placed two NTC's parallel and then two of this sets in series. The total resistance would than be equal to one lone NTC. For example 1 NTC is 100 ohms, two of them parallel would give 50 ohms, but if you set two of this parallel sets in series the sum is again 100 ohms. In this case setting NTC's parallel is no problem, or am I wrong? Aha that is a good point, the absorbed energy in the capacitors has no where to go. So actually how I mounted them now is nog dangerous, but not really functional?