The other ACA unit measure identical. 1.46 - 1.47 A when "bias" set to 12V and current decreases a bit during warm up.
Distortion this time measured at 1 kHz (8.2 and 4.1 ohm load). Both has 2nd harmonic as most dominant but very little at 8.2 ohm load. Output power is approx. 3.4 W at 8.2 ohm and 4.1 W at 4.1 ohm. I decreased the input signal a bit from 8.2 to 4.1 ohm load.
I think quite good results for both units. Then I can begin building the linear PSU's…..
Distortion this time measured at 1 kHz (8.2 and 4.1 ohm load). Both has 2nd harmonic as most dominant but very little at 8.2 ohm load. Output power is approx. 3.4 W at 8.2 ohm and 4.1 W at 4.1 ohm. I decreased the input signal a bit from 8.2 to 4.1 ohm load.
I think quite good results for both units. Then I can begin building the linear PSU's…..
Attachments
I've got one Amp Camp Amp V1.6 on order. My speakers are 4 Ohm 86dB. So I'd like to know how (/if) the Polk Audio LSi25 4 ohm 88 db sounds with your amp.
In the case of First Watt, I initially bought 2,000 sets of power supply
parts, and I am not quite done with them.
In general, noise, availability, and reliability of switchers are a problem.
The kind of switching supplies I would want for PL and FW are really quite
expensive, and would generally have to be custom made.
With regards to ACA and some other projects, a decent switcher has
the tremendous advantage of being simple and SAFE. The Amp Camp
event would not have been possible without them. Also the cost was
very cheap, as there was (and still is) a glut in the market on 19V
switchers used for laptops.
Thank you for the response. So if cost and convenience is not a problem the old type PSU is still the way to go for best performance.
The reliability could be a problem. 20 years may be a long time for a switcher.
I think these are cheap:
Elco low voltage 47000uF-63V - Elektrodump
Was Rifa in the old days…….
These Ecaps are good for linear PSU use……….
Hello,
I didnt go for these caps because they are rather big and i am not sure if one should use 63 volts caps if the actual voltage during use will be well under 30 volts.
I am using 40 volts models so for the same seize i will have more µF.
This store usually has professional gear. I like the termination on these caps having an oval terminal so one could use a wrench to prevent the terminal turning when you put some force on the screw.
A set of transformers ( when double mono) and a few Kemet caps from this shop and you will have some money left to get some Lundahl LL2733 chokes like my friend Nagini in vietnam is using for his ACA amp.
As you can see in my picture i am using 3,5 A rated diodes but they have no problem if i present a 2 A load because the choke will give the supply a steady non pulsating load even if there is around 70000µF in the supply and a big transformer.
Greetings, Eduard
I didnt go for these caps because they are rather big and i am not sure if one should use 63 volts caps if the actual voltage during use will be well under 30 volts.
I am using 40 volts models so for the same seize i will have more µF.
This store usually has professional gear. I like the termination on these caps having an oval terminal so one could use a wrench to prevent the terminal turning when you put some force on the screw.
A set of transformers ( when double mono) and a few Kemet caps from this shop and you will have some money left to get some Lundahl LL2733 chokes like my friend Nagini in vietnam is using for his ACA amp.
As you can see in my picture i am using 3,5 A rated diodes but they have no problem if i present a 2 A load because the choke will give the supply a steady non pulsating load even if there is around 70000µF in the supply and a big transformer.
Greetings, Eduard
Just finished up an ACA v1.5 - relatively easy build (good thing, since I haven't dusted off the soldering iron in a while) - everything worked fine on power up & bias adjusted nicely (although it took a while).
Did a little listening - replaced a crown pro amp between a Bottlehead Smash preamp & a pair of Klipsch Cornwalls. Very, very nice!! Lots of clarity & good bass response. I think this amp will stay in play.
I think this is clearly a case where simplicity in design can yield great results (if done carefully).
Rich
Did a little listening - replaced a crown pro amp between a Bottlehead Smash preamp & a pair of Klipsch Cornwalls. Very, very nice!! Lots of clarity & good bass response. I think this amp will stay in play.
I think this is clearly a case where simplicity in design can yield great results (if done carefully).
Rich
Yes, the amps are very predictable in behavior and only one adjustment. I made an effort to match components for each amp. I used 0.1% for the NFB resistors to be sure to have identical gain. I just looked at the Bourns blue pots to see which position they ended up in. Pretty close for the two ACAs?
Attachments
Ok....I was also assuming that the DC characteristic of the JFET was in the area where Ids will give a linear response to the input voltage. So Id will be something between 0 and Idss…...in the middle?
If I look at the characteristic of a N-channel JFET we are in the ohmic region? From the schematic I should be able to calculate Vg…..Vgs is always negative?
Hello,
We are all waiting for Meper to compare his new to build LCLC power supply with the '' lab power supply '' he is using now to do his measurements. Nelson Pass surely did enough measurements.
My Vietnamese friend really was surprised by the improvement the double coil choke did give. Using one coil on each side of the output from the bridge to the first cap. Meper will do the same because he has chokes with two windings as well but his chokes have more mH and less DCR.
In my home cinema set i am using an integrated class D ( Virtue sensation) I started with a LCLC supply using 100mH chokes, then switched to 2 sealed BIG batteries in series ( which start losing their charge every month a little quicker) So it was a new set of batteries or buying a set of BIG chokes from a Belgian member here , using an LL2733 ( that i got for cheap because it was the wrong current rating when i ordered it i was offered to me to keep it at a cheaper price, buy a surplus power transformer and some Kemet caps at mouser. Dont need huge values because the current being drawn isnt big.
I dont know about the psrr regarding this amp. It could be better than the ACA because that is very minimal design. The new power supply is cheaper than a set of batteries and it sounds wonderful.
A pity people dont often use choke input power supplies now that filament chokes with enough current rating are available.
greetings, eduard
We are all waiting for Meper to compare his new to build LCLC power supply with the '' lab power supply '' he is using now to do his measurements. Nelson Pass surely did enough measurements.
My Vietnamese friend really was surprised by the improvement the double coil choke did give. Using one coil on each side of the output from the bridge to the first cap. Meper will do the same because he has chokes with two windings as well but his chokes have more mH and less DCR.
In my home cinema set i am using an integrated class D ( Virtue sensation) I started with a LCLC supply using 100mH chokes, then switched to 2 sealed BIG batteries in series ( which start losing their charge every month a little quicker) So it was a new set of batteries or buying a set of BIG chokes from a Belgian member here , using an LL2733 ( that i got for cheap because it was the wrong current rating when i ordered it i was offered to me to keep it at a cheaper price, buy a surplus power transformer and some Kemet caps at mouser. Dont need huge values because the current being drawn isnt big.
I dont know about the psrr regarding this amp. It could be better than the ACA because that is very minimal design. The new power supply is cheaper than a set of batteries and it sounds wonderful.
A pity people dont often use choke input power supplies now that filament chokes with enough current rating are available.
greetings, eduard
It will take some time to finish the two PSUs for the ACA. Today I cut two boards from a 1m x 1m 25mm black MDF board I had in spare from a speaker project. When you live in a normal apartment it is a challenge to cut such a board also considering it is not every day I use a circular saw. One problem is all the dust. Circular saws needs to be handled with respect. They are quite powerful. But I got one step further…….I have two boards now. Large enough so chokes can be placed so they don't "talk" to each other.
The lap supply was just for testing the ACAs. It is a good feeling that you can slowly turn up the voltage and the current limiter is also nice. It has a fan built it so not good for listening test. Will wait for the LCLC PSUs to be finished before I will listen to the ACAs. So I will never know how the sound is compared to a laptop supply (I have the 19V's from the kit but they will just stay where they are). If someone wants them…...they can come and pick them up
The lap supply was just for testing the ACAs. It is a good feeling that you can slowly turn up the voltage and the current limiter is also nice. It has a fan built it so not good for listening test. Will wait for the LCLC PSUs to be finished before I will listen to the ACAs. So I will never know how the sound is compared to a laptop supply (I have the 19V's from the kit but they will just stay where they are). If someone wants them…...they can come and pick them up
Attachments
Ah, great, you're discussing JFET's again . Let me chime in...
While waiting for some other parts I had to order for a pair of ACAs, I've been thinking about what kind of cheap and in-production JFETs I could use instead of those LSK's and thus avoid overseas shipping.
@Loudthud wrote the following in Post #761:
So I've put together the little test circuit shown below and tested one specimen of every kind of JFET I had in the parts bin. Only with the BF246C I was unable to set a current of 5mA, so it seems to me that this part will not work as a substitute. All the other parts permitted the setting of 5mA through the 1k resistor (R9 in the original schematic), albeit with VGS not always in between -1V to -4V; the 2N5484 even showed a positive VGS.
Since I don't have a working ACA just yet, I can't make any real tests but only assume which parts would work OK. That would be the BF246B, BF245C and 2SK246BL. I'm not sure about the BF245B and the 2N5484. The BF246C will certainly not work, because the current through it is too high and thus the optimal bias point cannot be set.
After running the specimens through the test circuit, I tested each with a DCA75 for their VGSoff, gfs and ID values. All my devices have a significantly lower gfs than the typical 22 mA/V of the 2SK170. How would that influence the original circuit? Lower open-loop gain and thus higher distorion? I'm not that familiar with JFETs, but I'm trying to get my head around it.
. Let me chime in...While waiting for some other parts I had to order for a pair of ACAs, I've been thinking about what kind of cheap and in-production JFETs I could use instead of those LSK's and thus avoid overseas shipping.
@Loudthud wrote the following in Post #761:
So I've put together the little test circuit shown below and tested one specimen of every kind of JFET I had in the parts bin. Only with the BF246C I was unable to set a current of 5mA, so it seems to me that this part will not work as a substitute. All the other parts permitted the setting of 5mA through the 1k resistor (R9 in the original schematic), albeit with VGS not always in between -1V to -4V; the 2N5484 even showed a positive VGS.
Since I don't have a working ACA just yet, I can't make any real tests but only assume which parts would work OK. That would be the BF246B, BF245C and 2SK246BL. I'm not sure about the BF245B and the 2N5484. The BF246C will certainly not work, because the current through it is too high and thus the optimal bias point cannot be set.
After running the specimens through the test circuit, I tested each with a DCA75 for their VGSoff, gfs and ID values. All my devices have a significantly lower gfs than the typical 22 mA/V of the 2SK170. How would that influence the original circuit? Lower open-loop gain and thus higher distorion? I'm not that familiar with JFETs, but I'm trying to get my head around it
.Attachments
wrong ;
in ACA , JFet situation is like this : job of it is to keep steady voltage (sag) across R9 ; there need to be , say, 3 to 5V , usually aroundish 4V , to open Q1 properly
while that is the case , gate of JFet can be set anything from ( observe P1) up from GND potential to U/2 potential
so , mentioned 4V(ish) voltage is attached to Q1 , while having nothing with possible or needed Ugs voltage for JFet
what is most important here is that you can bias input JFet in a way to have ~4mA Iq , while still having decent xconductance - read - Rin , cojones, whatever ...... to fulfill all expectations regarding pumping that input capacitance of Q1 , augmented by Herr Mueller's effect.
so - choice of possible JFets is broad , even broader if you insert appropriate resistor between top of R7 and top of JFet drain line , thus slightly decreasing JFet drain voltage ....... all with JFet sane dissipation in mind
in ACA , JFet situation is like this : job of it is to keep steady voltage (sag) across R9 ; there need to be , say, 3 to 5V , usually aroundish 4V , to open Q1 properly
while that is the case , gate of JFet can be set anything from ( observe P1) up from GND potential to U/2 potential
so , mentioned 4V(ish) voltage is attached to Q1 , while having nothing with possible or needed Ugs voltage for JFet
what is most important here is that you can bias input JFet in a way to have ~4mA Iq , while still having decent xconductance - read - Rin , cojones, whatever ...... to fulfill all expectations regarding pumping that input capacitance of Q1 , augmented by Herr Mueller's effect.
so - choice of possible JFets is broad , even broader if you insert appropriate resistor between top of R7 and top of JFet drain line , thus slightly decreasing JFet drain voltage ....... all with JFet sane dissipation in mind
Last edited: