thanks Rob, I was wasn't sure what they are called, I guess a search for fuse would have got me there🙂
Has anyone that was expecting pcb's not received them yet? I would like to know so I can take care of one more piece of business and start soldering.
Thanks Tad
Thanks Tad
emitter resistors
I can offer this:
http://www.diyaudio.com/forums/solid-state/52459-smaller-leach-amp-v1.html
Go to post #881 (Jacco) and read through #909 (Pooge) for an overview. It comes down to "how big is the bulls eye?". Availability, cost, fine tuning of the circuit all come into play...along with "can I hear it?" for myself.
Lower inductance is favored. Non-inductance depends upon personal tastes.
HTH
Maybe...& I'm not just trying to be cute, I don't know.
I can offer this:
http://www.diyaudio.com/forums/solid-state/52459-smaller-leach-amp-v1.html
Go to post #881 (Jacco) and read through #909 (Pooge) for an overview. It comes down to "how big is the bulls eye?". Availability, cost, fine tuning of the circuit all come into play...along with "can I hear it?" for myself.
Lower inductance is favored. Non-inductance depends upon personal tastes.
HTH
Last edited:
yeah well I dont have a clue, just using some math. On paper inductance free is looking better🙂
I guess theres no right or wrong, just peoples experiences.
Last edited:
I would not get over paranoid about the inductance of a .33 ohm resistor.
I WOULD be more concerned about magnetic material in the LR resistor in the output if the coil is near, or wrapped around, the resistor. This coil is very desirably an air coil. Having magnetic material inside may saturate under high output currents.
Be that as it may, the emitter resistors cary a lot of high and variable current. They are local feedback resistors to the output transistors, and are where the signals originate for the protection circuit. So in theory, lower inductance would seem to be "better", but lower TCR MAY be more important. The thing is, you can get all the parameters more better in a more better resistor, and get more reliable construction.
That being said, there has been no indication from Leach that cheap sand cast resistors will cause problems with the amp. I'm sure he, any many other constructors, have used these cheap parts and lost no sleep. Whether or not there is any perceivable sound difference, I haven't compared.
I WOULD be more concerned about magnetic material in the LR resistor in the output if the coil is near, or wrapped around, the resistor. This coil is very desirably an air coil. Having magnetic material inside may saturate under high output currents.
Be that as it may, the emitter resistors cary a lot of high and variable current. They are local feedback resistors to the output transistors, and are where the signals originate for the protection circuit. So in theory, lower inductance would seem to be "better", but lower TCR MAY be more important. The thing is, you can get all the parameters more better in a more better resistor, and get more reliable construction.
That being said, there has been no indication from Leach that cheap sand cast resistors will cause problems with the amp. I'm sure he, any many other constructors, have used these cheap parts and lost no sleep. Whether or not there is any perceivable sound difference, I haven't compared.
The output resistors
As pooge and others have said Dr. Leach makes no comment as to wether the output resistors are inductive or not. This being said I may choose to incorporate the lower cost sand cast items in my build. Thanks. Money being what it is has a large voice in this project.
There seems to be enough room to use just about any resistor type one would like.
I never claimed to be an audiophile just a hobbyist. I spend prodigious amounts of time on the chassis and outer case. That is what the wife has to see.
Tad
As pooge and others have said Dr. Leach makes no comment as to wether the output resistors are inductive or not. This being said I may choose to incorporate the lower cost sand cast items in my build. Thanks. Money being what it is has a large voice in this project.
There seems to be enough room to use just about any resistor type one would like.
I never claimed to be an audiophile just a hobbyist. I spend prodigious amounts of time on the chassis and outer case. That is what the wife has to see.
Tad
Parallel resistor calculator
For those interested in a tighter tolerance on Re
Parallel Resistance Calculator R1 + R2 = equivalent resistor R equiv total resistor finder - sengpielaudio Sengpiel Berlin
For those interested in a tighter tolerance on Re
Parallel Resistance Calculator R1 + R2 = equivalent resistor R equiv total resistor finder - sengpielaudio Sengpiel Berlin
Tighter tolerance (1% or better) on the "better" emitter resistors is certainly something to consider and strive for, especially regarding current sharing among all of those output transistors.
low R measurement
Since the resistances we are concerned with are so small, I'm considering measuring V through them. Is a 9V battery sufficient as a source?
Since the resistances we are concerned with are so small, I'm considering measuring V through them. Is a 9V battery sufficient as a source?
Hi,
you don't need much current but you do need fairly consistent current.
I use a CCS based on a 317. It can be preset anywhere between 10mA and 500mA.
attach a series string of DUTs + one REF.
adjust the current until you have 190.0 to 195.0 mV drop across ref.
Measure the voltage across all the others and note them.
Replace the DUTs and start again ensuring that Vref is identical to the first Vref reading.
Once you have them all labelled you can select out your pairs or quads or whatever you need and string them together as a final check against Vref.
Most resistors with ~190mV across them will change temperature. You have to use a method that allows for this temperature change. Either keep them all at ambient (the same each time) or let them warm up to a fixed level above a fixed ambient. Both methods need consideration to remove the error.
Back to your question, a 9V battery may not be enough.
If you do this well, you can measure the lead out resistance quite accurately and you will discover that the distance between the probes along the lead out wires does seriously affect your readings.
you don't need much current but you do need fairly consistent current.
I use a CCS based on a 317. It can be preset anywhere between 10mA and 500mA.
attach a series string of DUTs + one REF.
adjust the current until you have 190.0 to 195.0 mV drop across ref.
Measure the voltage across all the others and note them.
Replace the DUTs and start again ensuring that Vref is identical to the first Vref reading.
Once you have them all labelled you can select out your pairs or quads or whatever you need and string them together as a final check against Vref.
Most resistors with ~190mV across them will change temperature. You have to use a method that allows for this temperature change. Either keep them all at ambient (the same each time) or let them warm up to a fixed level above a fixed ambient. Both methods need consideration to remove the error.
Back to your question, a 9V battery may not be enough.
If you do this well, you can measure the lead out resistance quite accurately and you will discover that the distance between the probes along the lead out wires does seriously affect your readings.
Last edited:
Thanks Pooge, I like to avoid the mythology surrounding hifi so I will take that on board when I make my order. One point that does make sense, is the more expensive ones are 1%, at greater voltage swings and power ratings we have in this pcb, it may be nice to know current sharing is optimised.
I find the idea of buying lots of resistors and matching a bit of a false economy, unless you happen to have several hundred lying around.
I find the idea of buying lots of resistors and matching a bit of a false economy, unless you happen to have several hundred lying around.
I should qualify my remarks about power sharing. There will be much more variance in the transistor parameters than the emitter resistors, and therefore a benefit to matching same sex outputs. As to the sensitivity of current sharing due to variances of the transistors vs. that of the emitter resistors, I am not in a position to say. However, whatever difference there is in the transistors, the emitter resistors, via feedback, help to reduce the current differences. I do not recall ever reading anything about the issue of variance of transistor parameters vs. variance of transistor parameters as a sensitivity analysis, other than that Self, for example, indicates that the higher the Re, the better the current sharing (but the more the distortion, due to compression), and found, if I recall, that .22 ohm resistors seemed to be at a happy medium. Since Leach specifies .33 ohm, and Jens has used .47 ohm, these values would indicate a tendency towards the current sharing side of .22 ohms.
Power Supplies
Hello all,
I have crafted a generalized power supply board that can take up to 8 caps (ie: 4 per rail). The PCB is set up for either 40 or 35 mm cans with 2 or 4 pin snap connectors. For the prototype I will be using 8 Panasonic ECE-T2AP103FA caps. That amounts to a total of 40,000uF per rail at 100 volts. The PCB also has places for bleeder resistors, 2 diode bridges and screw connectors to bind to the outside world.
The PCBs will be coming from fab within the next few days. They measure 3.5 x 8.5 inches and are 2.4mm thick, double sided with 2oz copper. They have soldermask on both side and silkscreen on the top. I will post a PIC once I get the PCBs from fab. That should be within a few days.
Do these interest anyone? I would be willing to front a group buy if their is sufficient interest in them. I would expect cost to be between $10 to $12 each (plus shipping) dependent up the number that I order from the fab shop.
This is a general purpose power supply board that would work with most amplifiers and not just the Leach. I've also got a smaller version with only 4 caps (ie: 2 per rail).
Let me know if either of these are interesting.
Hello all,
I have crafted a generalized power supply board that can take up to 8 caps (ie: 4 per rail). The PCB is set up for either 40 or 35 mm cans with 2 or 4 pin snap connectors. For the prototype I will be using 8 Panasonic ECE-T2AP103FA caps. That amounts to a total of 40,000uF per rail at 100 volts. The PCB also has places for bleeder resistors, 2 diode bridges and screw connectors to bind to the outside world.
The PCBs will be coming from fab within the next few days. They measure 3.5 x 8.5 inches and are 2.4mm thick, double sided with 2oz copper. They have soldermask on both side and silkscreen on the top. I will post a PIC once I get the PCBs from fab. That should be within a few days.
Do these interest anyone? I would be willing to front a group buy if their is sufficient interest in them. I would expect cost to be between $10 to $12 each (plus shipping) dependent up the number that I order from the fab shop.
This is a general purpose power supply board that would work with most amplifiers and not just the Leach. I've also got a smaller version with only 4 caps (ie: 2 per rail).
Let me know if either of these are interesting.
a stereo pair of 12output PCBs will need 24 resistors.
Don't buy extra unless you want to.
All you have to do is group them into four 6piece groups. You will probably find that three of your groups will each be within 0.75% and the worst group probably within 1.5%.
You can mark them with an indelible pen so when ever you measure you can see what correction to apply to find the device Ic
Carl,
I have always liked the neat well thoughtout aspect of mounting the big electros on a board. You can tentatively count me in on this. I would like to see some pics.
Is that two bridge rectifiers per rail or side?
You are going to run 100 volt rails? Should be one monster. Mono or stereo chassis?
Tad
I have always liked the neat well thoughtout aspect of mounting the big electros on a board. You can tentatively count me in on this. I would like to see some pics.
Is that two bridge rectifiers per rail or side?
You are going to run 100 volt rails? Should be one monster. Mono or stereo chassis?
Tad
Tad,
I expect to receive the proto boards early next week. I will immediately build one up and test it (to verify that I didn't do something stupid on the layout), shoot some PICs and post them here. That should happen by this time next week.
As to my build, my first pair will be monoblocks. Reliability is of course a tradeoff between voltage and load. I'll probably settle on rail voltage between 80 and 90 volts. At some point after I have lived with the amps for a while I would like to experiment with voltage regulating the frontend. I most likely will adapt the Bob Ellis/Jens Rasmussen voltage regulator for that.
I expect to receive the proto boards early next week. I will immediately build one up and test it (to verify that I didn't do something stupid on the layout), shoot some PICs and post them here. That should happen by this time next week.
As to my build, my first pair will be monoblocks. Reliability is of course a tradeoff between voltage and load. I'll probably settle on rail voltage between 80 and 90 volts. At some point after I have lived with the amps for a while I would like to experiment with voltage regulating the frontend. I most likely will adapt the Bob Ellis/Jens Rasmussen voltage regulator for that.
Re
Its looking like I'll do something like AndrewT for my emitter resistors. Key to this is a source selling the Yageo 0.51 ohm wire wound from a bin for cheap...$0.05 ea.
I'm looking at some 5% 2W metal oxide film in values of 2 & 2.7 ohms to use in combination.
I'll be making my purchase of the Yageo on Friday, April 23. I would be glad to purchase additional and send them along, at cost.
Its looking like I'll do something like AndrewT for my emitter resistors. Key to this is a source selling the Yageo 0.51 ohm wire wound from a bin for cheap...$0.05 ea.
I'm looking at some 5% 2W metal oxide film in values of 2 & 2.7 ohms to use in combination.
I'll be making my purchase of the Yageo on Friday, April 23. I would be glad to purchase additional and send them along, at cost.
Any more suggestions on the driver and predriver transistors. I have found the 2sa1837/sc4793 at Ampslab for use as a predriver. Vceo-230 v, 1 A, Ft100 and Hfe 100-320, The footprint is exactly reverse which makes things easier. Any ideas.
Now for a very good driver? The mje15030 has excellent Vceo buts lacks speed. Any suggestions on how to determine the driver load with 12 mje3281 outputs. That information would make selection easier.
And thanks for the link to the Vas driver transistor thread. Much useful info there.
Tad
Now for a very good driver? The mje15030 has excellent Vceo buts lacks speed. Any suggestions on how to determine the driver load with 12 mje3281 outputs. That information would make selection easier.
And thanks for the link to the Vas driver transistor thread. Much useful info there.
Tad
- Status
- Not open for further replies.