Terry
so now the fiddling and experimenting starts
It is always a bitter sweet victory when one completes an amp that performs as well as the Leachmussen. There is always the feeling that you could have done something better and it nags at you while listen to the music. Put these thoughts aside for now - but not too long - we want the answers
Arius -
I am interested to have more details on your P/S solution. I would imagine that integrating the DC Detection circuitry that Tony is working on would not be too difficult.
Jens
You are getting slack, or is that slap Nice looking boards. I have a few questions if you don't mind.
1. What are the emmitor resistors - they don't look like standard wirewounds.
2. I notice the caps that are being used are much larger than the 15000uF Cornel Dublier's that some are using. Is this indicative of better quality? Does anybody have any comments with regards to this.
so now the fiddling and experimenting starts
It is always a bitter sweet victory when one completes an amp that performs as well as the Leachmussen. There is always the feeling that you could have done something better and it nags at you while listen to the music. Put these thoughts aside for now - but not too long - we want the answers
Arius -
I am interested to have more details on your P/S solution. I would imagine that integrating the DC Detection circuitry that Tony is working on would not be too difficult.
Jens
You are getting slack, or is that slap
Nice looking boards. I have a few questions if you don't mind. 1. What are the emmitor resistors - they don't look like standard wirewounds.
2. I notice the caps that are being used are much larger than the 15000uF Cornel Dublier's that some are using. Is this indicative of better quality? Does anybody have any comments with regards to this.
still4given said:
Yes, the outside legs (from the speaker output ends) of an opposisng pair.
If that is hard to measure, just up the bias to at least 150mA to make it roughly equivalent to Leach's recommended bias, since you have an extra pair of outputs. Leach uses .33ohm resistors, so if you used .47 ohm, the Self optimum bias will not be drastically different than Leach's. You should top out around 177mA or so for Self's optimum bias, which has the lowest variation in output impedance through the crossover region. You could try 175mA or so to start, measured as you've done already, through removed fuse holder. This will be close enough. Self found that distortion went up if you go higher into AB. I wouldn't go higher than Self's value unless trying for class A, especially with your present transformer.
I was thinking more of trying something in the line of 1 watt in class A.
Computer grade capacitors are generally a lot bigger in size.
Long Life and 105C types have bigger dimensions too.
Sikorel's are still generally seen as among the absolute best electrolytics.
A 10.000uF/63V measures 2"x4" , a 10.000uF/100V is 3"x4", much bigger than general use Epcos caps.
Still, some brands used to make specialty small dimension capacitors of high quality, like Roederstein with the EY series.
There is a correlation between dimensions, capacity, and voltage.
Something like 15 years ago i found a decent one using dimension analysis, by comparing a number of different size capacitors of the same series.
I have not seen tech articles on the size subject on Cornell's pages yet, however i have read some pdf's on new technology capacitors, dowloaded from Vishay and Epcos.
Computer grade capacitors are generally a lot bigger in size.
Long Life and 105C types have bigger dimensions too.
Sikorel's are still generally seen as among the absolute best electrolytics.
A 10.000uF/63V measures 2"x4" , a 10.000uF/100V is 3"x4", much bigger than general use Epcos caps.
Still, some brands used to make specialty small dimension capacitors of high quality, like Roederstein with the EY series.
There is a correlation between dimensions, capacity, and voltage.
Something like 15 years ago i found a decent one using dimension analysis, by comparing a number of different size capacitors of the same series.
I have not seen tech articles on the size subject on Cornell's pages yet, however i have read some pdf's on new technology capacitors, dowloaded from Vishay and Epcos.
Byrd said:
Jens
You are getting slack, or is that slap
1. What are the emmitor resistors - they don't look like standard wirewounds.
2. I notice the caps that are being used are much larger than the 15000uF Cornel Dublier's that some are using. Is this indicative of better quality? Does anybody have any comments with regards to this.
1) They are the cheapest I could find here in Denmark
2) They are 10.000uF 100V types from Cornel Dublier's..... Generally bigger means longer life when it comes to caps.
\Jens
Hi Guys,
OK, I upped the current to 150mA. Bo I learned a lesson about connecting the meter leads without completely discharging the caps. 
Rigged up a light bulb to drain them from now on.
It seems to have smoothed out the highs a touch. It is noticeably smoother at higher gain now. The snare drum hits were a tiny bit harsh before. Now they are very real sounding.
I have been listening through two pair of JBLs this morning. They measure 3.3ohms hooked in parallel. The heatsinks are just barely warm.
I don't think I'm interested in taking this amp into class A. That Krell has left a bad taste in my mouth. It will probably be good to heat the garage this winter.
With this Low TIM, I can close my eyes and tell were each instrument is on stage. I can even tell which ones are farther away. That's wonderful.
For the record, I am using a Carver PDR-10 CDR to play the CDs. Taking the digital out to a Yamaha CX-1000U digital in and using the Source Direct setting. About as uninterrupted a signal as I can get. No EQ of any sort to mess things up. Listening to Pheonix Jazz Quartet, Kenny G, Joey DeFrancesco with Jimmy Smith, Michael English, Peter Cetera, Stevie Wonder, Chicago, etc.
All sound as they should IMO.
Blessings, Terry
OK, I upped the current to 150mA. Bo I learned a lesson about connecting the meter leads without completely discharging the caps.
Rigged up a light bulb to drain them from now on.
It seems to have smoothed out the highs a touch. It is noticeably smoother at higher gain now. The snare drum hits were a tiny bit harsh before. Now they are very real sounding.
I have been listening through two pair of JBLs this morning. They measure 3.3ohms hooked in parallel. The heatsinks are just barely warm.
I don't think I'm interested in taking this amp into class A. That Krell has left a bad taste in my mouth. It will probably be good to heat the garage this winter.
With this Low TIM, I can close my eyes and tell were each instrument is on stage. I can even tell which ones are farther away. That's wonderful.
For the record, I am using a Carver PDR-10 CDR to play the CDs. Taking the digital out to a Yamaha CX-1000U digital in and using the Source Direct setting. About as uninterrupted a signal as I can get. No EQ of any sort to mess things up. Listening to Pheonix Jazz Quartet, Kenny G, Joey DeFrancesco with Jimmy Smith, Michael English, Peter Cetera, Stevie Wonder, Chicago, etc.
All sound as they should IMO.
Blessings, Terry
Hi Terry,
do you remember what I said about SOAR on your higher voltage Leach clone?
4 ohm moderate speakers are the low limit if you want long life out of your transistors.
8 ohm severe speakers can be driven all day to any level short of clipping provided the heatsinks stay cool.
1.65 ohm is just stretching things - keep the volume well down.
do you remember what I said about SOAR on your higher voltage Leach clone?
4 ohm moderate speakers are the low limit if you want long life out of your transistors.
8 ohm severe speakers can be driven all day to any level short of clipping provided the heatsinks stay cool.
1.65 ohm is just stretching things - keep the volume well down.
still4given said:
OK, I upped the current to 150mA. Bo I learned a lesson about connecting the meter leads without completely discharging the caps.
Rigged up a light bulb to drain them from now on.
This is one of the reasons why I've been harping the voltage measurement method!
Just get some tip jacks like the ones shown in Figure 9 of this page:
http://dkc3.digikey.com/PDF/T052/0250.pdf
Mount them on the chassis so you can insert your DVM probe tips, and monitor the bias that way.
These particular jacks mount from the outside of the case, so they would have to be disconnected from the boards to remove them. I haven't found any that mount from the inside so they can be easily removed from the case along with the boards. However, the connection between the emitter resistors and the probe jacks is not signal line, so cheap inline connectors could be used to disconnect the boards from the tip jacks to make it quick and easy to remove the boards from the case.
pooge said:
Hi pooge,
I let the amp pay for a couple of hours and then unplugged the speakers and inputs and rechecked using the voltage across the emitters this time. The left channel was only 40mV and the right was 70mV. Strange. No wonder the right channel felt warmer than the left. I reset them using the voltage method this time. I will check again later.
I'm wondering about installing those jacks you listed. How often do you reset the bias on your amps. Seems like after you get it setup you wouldn't need to do that any more. Also, don't you still need to open the amp to adjust it anyway?
Thanks again, Terry
Terry,
Sounds like you didn't take the time to let the bias stop drifting when you biased it. Those readings are way off. I would think it would take at least an hour to be comfortably stabilized, to make the heat sinks stabilize at a steady temp.
Also, disconnect the inputs so you don't get a DC offset when you adjust. When you adjust, adjust the value half way between your present reading and the desired reading, wait for it to stabilize a bit, and keep adjusting half way until you reach a steady value for a good while. At least 15-20 minutes or so. That way, you won't be overshooting, then undershooting again and again.
It's not that I bias often, but I would bias after each change in components. Also, I think it would be safer, as you have found. Also, from what I've read, the voltage method is arguably more accurate, and using the pin jacks is more safe than trying to jam a probe down into the amp.
Also, I'm starting to be of the opinion that measuring the bias in the amp's operating environment (i.e., all closed up) is preferable. When the amp is opened up, it can run a lot cooler. Also, it can be more "breezy" around the components. This seems to cause drifting in the bias, especially the current. The voltage seems to stay in a tighter window. Just put your finger on a transistor, even the small signal ones, and watch the bias current drift. I found that heat sinking the VAS and predriver lowered the drift a lot in my admittedly stuffed enclosure.
In most cases you have to open the top to adjust the bias, unless you've designed for it otherwise with a chassis mounted pot, like I suggested before. However, if you lift the top, adjust half way, and put the top back on until the temps and bias stabilize again before adjusting again, you can make do.
Take some time to do the adjustment. Your large heat sinks will take time to adjust. I don't think the ten minutes Leach suggested is anywhere close to enough.
Sounds like you didn't take the time to let the bias stop drifting when you biased it. Those readings are way off. I would think it would take at least an hour to be comfortably stabilized, to make the heat sinks stabilize at a steady temp.
Also, disconnect the inputs so you don't get a DC offset when you adjust. When you adjust, adjust the value half way between your present reading and the desired reading, wait for it to stabilize a bit, and keep adjusting half way until you reach a steady value for a good while. At least 15-20 minutes or so. That way, you won't be overshooting, then undershooting again and again.
It's not that I bias often, but I would bias after each change in components. Also, I think it would be safer, as you have found. Also, from what I've read, the voltage method is arguably more accurate, and using the pin jacks is more safe than trying to jam a probe down into the amp.
Also, I'm starting to be of the opinion that measuring the bias in the amp's operating environment (i.e., all closed up) is preferable. When the amp is opened up, it can run a lot cooler. Also, it can be more "breezy" around the components. This seems to cause drifting in the bias, especially the current. The voltage seems to stay in a tighter window. Just put your finger on a transistor, even the small signal ones, and watch the bias current drift. I found that heat sinking the VAS and predriver lowered the drift a lot in my admittedly stuffed enclosure.
In most cases you have to open the top to adjust the bias, unless you've designed for it otherwise with a chassis mounted pot, like I suggested before. However, if you lift the top, adjust half way, and put the top back on until the temps and bias stabilize again before adjusting again, you can make do.
Take some time to do the adjustment. Your large heat sinks will take time to adjust. I don't think the ten minutes Leach suggested is anywhere close to enough.
Hi Terry,
Glad things worked out well for you. I wouldn't worry too much about the bias being exact. Pooge's advice is right on the money. I would heatsink my VAS and predriver but I'll probably be too lazy to bring the Vq points out to chassis. Those are high voltage points not covered by the protection circuit.
I'll take your success to inspire me to hurry along with my own amplifier. I just can't seem to make up my mind with the various tweaks possible. Haha. One confirmed tweak is to use OnSemi's Thermal Trak 5-pin devices. Waiting on the samples before I start layout.
Byrd expressed some interest in my ongoing AC Soft-Start project. I'll open a new thread for it in a few weeks time. I work really slow, sorry.
Happy Listening...
Glad things worked out well for you. I wouldn't worry too much about the bias being exact. Pooge's advice is right on the money. I would heatsink my VAS and predriver but I'll probably be too lazy to bring the Vq points out to chassis. Those are high voltage points not covered by the protection circuit.
I'll take your success to inspire me to hurry along with my own amplifier. I just can't seem to make up my mind with the various tweaks possible. Haha. One confirmed tweak is to use OnSemi's Thermal Trak 5-pin devices. Waiting on the samples before I start layout.
Byrd expressed some interest in my ongoing AC Soft-Start project. I'll open a new thread for it in a few weeks time. I work really slow, sorry.
Happy Listening...
Arius said:
I can understand about being too lazy to bring the Vq points out to the chassis. But there is little difference in voltage from the speaker outputs brought out to chassis, and certainly not high voltage. The protection circuit is an issue, but not much if you don't get too stupid about it. The tip jacks are too small to receive a banana plug or an RCA, so it wouldn't worry me. I'd worry more about starting the amp up with the fuse pulled out to check bias current, or sticking a probe into the belly of the amp. Sh*t happens!! It would be a lot harder to get up the desire to install the jacks after the amp is up and running, though, so planning to do it up front would be much easier.
I got really excited when I saw those Thermal Trak devices. Seemed like a gift from above. However, after the initial excitement wore off, I realized that Self's studies on thermal tracking found overcompensation when he put the temperature sensor on the case of a TO-3 package, and the Thermal Traks would respond even faster. He ended up finding better tracking with some thermal insulation in between. I had trouble following his British English at times. However, I assume the problem with "instant" measurement is that it can react too fast when transients come through, heat up quickly, then die off. The bias maybe reacts too quickly when it doesn't have to, because the transients are gone so quickly. Perhaps the Thermal Trak devices would be the Cat's Meow if some kind of bias tracking circuit was specifically designed to operate with them. Such a circuit is probably not far off. It will probably integrate the temperate in some fashion to mimic a delay like Self found useful with his insulator. Self's experiment was with a T0-3 package. The Thermal Trak should respond even faster. I did notice, though, that the thermal resistance to the case of a plastic output transistor is similar to the thermal resistance Self found to be useful as a thermal resistance between a TO-3 case and a temperature sensor--so perhaps piggy-backing a temperature sensor on the back of a plastic output would work to perfection! You could even use the B-E junction of a spare output transistor that you set aside because it didn't match. If you do this, I would put one on top of every output transistor, even if you don't use them all as sensors, so that all of the output transistors work in the same thermal environment so they share current better.
Hi Pooge,
A thought came to me as to why my bias was off the first time I set it. The amp was sitting right next to my little window shaker. The right channel would have been running much cooler than the left due to the cool air hitting that side of the amp. When I set it the second time using voltage drop across the resistors, it was in the morning and I hadn't yet turned on the AC. Another lesson learned. It is my hope that my blunders are helping others to not make the same mistakes. I know I am learning from them. I may look into those little jacks. They should be easy enough to install. I wouldn't think they would need very heavy wire and attaching them to the resistor leads should very simple. I did have all leads disconnected when setting the bias and I did take over an hour to set it but I didn't do the "half way" method. I will try that the next time.
Blessings, Terry
A thought came to me as to why my bias was off the first time I set it. The amp was sitting right next to my little window shaker. The right channel would have been running much cooler than the left due to the cool air hitting that side of the amp. When I set it the second time using voltage drop across the resistors, it was in the morning and I hadn't yet turned on the AC. Another lesson learned. It is my hope that my blunders are helping others to not make the same mistakes. I know I am learning from them. I may look into those little jacks. They should be easy enough to install. I wouldn't think they would need very heavy wire and attaching them to the resistor leads should very simple. I did have all leads disconnected when setting the bias and I did take over an hour to set it but I didn't do the "half way" method. I will try that the next time.
Blessings, Terry
I'm looking to order some boards, and I wonder if I'll be able to use some power transistors I already have, without buying new ones for this project.
Would anyone know if the MJ21193/21194 pair, or the MJ15003/15004 pair can be used in place of the MJL4281/4302?
I'm willing to handle the different mounting/heatsinking needs of my TO3 devices, if they can fit the bill electrically.
Thanks
Adrian
Would anyone know if the MJ21193/21194 pair, or the MJ15003/15004 pair can be used in place of the MJL4281/4302?
I'm willing to handle the different mounting/heatsinking needs of my TO3 devices, if they can fit the bill electrically.
Thanks
Adrian
funberry said:
Have you checked out Dr. Leach's website?
I'm not sure you can order any more boards from this group buy. I thinks they are all gone, but I believe you can still get them from Dr. Leach. His original design is set up for TO3's. The purpose of this GB design was to take advantage of the flat pack transistors and to get rid of all the wires required to hook up TO3's. There have been many, many Low TIM amps built over the years using TO3's however and Dr. Leach's website if very thorough.
Yes it seems stable and it sounds very good. I will check it again in about a week and see if it is still running the same. I'm enjoying listening to it right now. Nothing like a new amp to inspire one to get out the "good stuff" and do some critical listening.
Blessings, Terry
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.