What is the point of having low distortion if it is buried below the noise floor.
Increasing hFE will usually lower the noise performance of individual transistor.
Jay,
Are you suggesting that the BC546 as used by Apex on so many designs may be just fine, or even better than 2SC1845 with regards to noise performance?
Are you suggesting that the BC546 as used by Apex on so many designs may be just fine, or even better than 2SC1845 with regards to noise performance?
No, 2SC1845 has so many advantages over BC546, even noise... It is one of the famous old transistor from NEC if I'm not mistaken. I haven't tried KSC1845 but other KSC transistors that I have tried compared favorably with the old stuffs and much better than the unpopular brand such as National...
Just remember the suffix that represents hFE classification...
Bimo modification even makes 1845 more suitable than BC546...
Oops!. Be careful with labels.
I am very relieved/pleased that everything works!. Thanks for posting.
Also A/B with original fx-8 and Bimo mod would be great!. Looking forward for it.
reg
Prasi
The plastic bag was labeled by vendor with a sharpie marker and the decimal was rubbed off. This is an excellent example of why we should use 2K2 vs 22K notation rather than 2.2K notation. Also, shows importance of double checking label vs color bands (drudgery) or using a DMM to check prior to installing.
Good thing no harm was done - although I had safety resistors in place and would have powered off at first sign of excessive heat or smell/smoke.
I think you want to match a bjt i/p diff pair for Vbe, something in the order of <10mV. It is easy enough to make a jig, supply the same bias as is used in the amp and measure across the collectors for a close match.
For the Latfets I suggest you use/try the Semelab parts like ALF16N16W/ALF16P16W, they work great as they are dual die with a 250W, 160 or 200V,16A ratings. With a LME49830 driver I measure around 600-650ns rise-fall times. Is that important, hard to say but it sure sounds good. you would be hard pressed to get a bjt o/p stage to slew that fast. Near perfect square waves.
For the Latfets I suggest you use/try the Semelab parts like ALF16N16W/ALF16P16W, they work great as they are dual die with a 250W, 160 or 200V,16A ratings. With a LME49830 driver I measure around 600-650ns rise-fall times. Is that important, hard to say but it sure sounds good. you would be hard pressed to get a bjt o/p stage to slew that fast. Near perfect square waves.
Last edited:
In this case, I will await for the 2SC1845 before I do any real serious listening or A/B comparison with the stock FX8, FH9, or JFET Circlotron.
Rsavas:
Is there a link for measurement rig to do this? Or is the rig essentially the IPS on a breadboard and have a DVM plugged in to monitor differential V across collectors for zero volts into base and keep popping transistors under test in until a match with low voltage differential is found? Will Hfe also be automatically matched or is that yet a second process to weed out?
Last edited:
Thanks for the layout! A great little amp. I am glad it sings too. Always trepidation when firing up for first time: first look for and avoid smells and magic smoke, then secondly look for the sign of first sound and then comes a sigh of relief that it all works as it should. 🙂
Happiness is a hot soldering iron...
I do want to note that I put in a Wima 1uF 63v MKP for the centrally located cap, and now have a low ESR 330uF 25v SEPF Panasonic OSCON for the feedback reservoir cap, all 470uF power rail caps are 63v Panasonics. I still need to make a 1-2uH output inductor. Any cheap (free scavenged) sources of 18ga magnet wires people can suggest?
Last edited:
Thanks for the layout! A great little amp. I am glad it sings too. Always trepidation when firing up for first time: first look for and avoid smells and magic smoke, then secondly look for the sign of first sound and then comes a sigh of relief that it all works as it should. 🙂
Happiness is a hot soldering iron...
I do want to note that I put in a Wima 1uF 63v MKP for the centrally located cap, and now have a low ESR 330uF 25v SEPF Panasonic OSCON for the feedback reservoir cap, all 470uF power rail caps are 63v Panasonics. I still need to make a 1-2uH output inductor. Any cheap (free scavenged) sources of 18ga magnet wires people can suggest?
thats great to know and i am really glad...
for inductor, i remember building my first inductor salvaged from a broken desktop SMPS.🙂
I broke down and just ordered a 50ft spool of *green* enameled 18ga copper magnet wire. 😛
Can't wait to get back and listen to this amp some more. This time with a bit more volume. 🙂
Can't wait to get back and listen to this amp some more. This time with a bit more volume. 🙂
FX8 Bimo Mod with matched 2SC1845's LTP
The 2SC1845's arrived today. Using a transistor tester, I found a pair with an exact Hfe match of 623 each. I installed them and bound them with shrink tube to be one thermal unit. I switched out R11 from 1k to 680R to get the same gain as my other amps. The new transistors are under the white shrink tube near the LED.
Listening to it now - sounds very very good. More time needed and some measurements of the speaker distortion will be in order but I will spend some time with different tracks to get sense. I am really enjoying its sound so far though.
Oh, and btw, the matched 2SC1845's now give me only a 1mV offset! (it was 60mV before with the randomly selected BC546's)
The 2SC1845's arrived today. Using a transistor tester, I found a pair with an exact Hfe match of 623 each. I installed them and bound them with shrink tube to be one thermal unit. I switched out R11 from 1k to 680R to get the same gain as my other amps. The new transistors are under the white shrink tube near the LED.
Listening to it now - sounds very very good. More time needed and some measurements of the speaker distortion will be in order but I will spend some time with different tracks to get sense. I am really enjoying its sound so far though.
Oh, and btw, the matched 2SC1845's now give me only a 1mV offset! (it was 60mV before with the randomly selected BC546's)
Attachments
So far, the FX8 Bimo mod appears to measure the best with my speakers in terms of lowest harmonic distortion (as measured with a microphone). The FX8 Bimo mod seems to have a slight 1.5dB depression in the response though from 2kHz to 5-7kHz as compared to the other amps which all track identically through that region. I wonder what can be causing this?
Measurements made with calibrated UMIK-1 (Cross Spectrum Labs calibrated), GRS Research XL-S speaker, all at 0.5m distance on tweeter axis with speaker on a 30in tall stand for tweeter 42in above floor. All amps used same Dx PSU with 35v rails and 250VA transformer.
Amps tested include: FX8 (5cm x 5cm version), FH9 (FX8 5cm modified with vertical FETs), Dx amp, and Inv JFET Circlophone.
Here is measured frequency response of speaker with different amps (FX8 Bimo is red curve):
Here is measured HD of speaker with FX8 5cm (note rising HD above 1kHz) - all values in legend are with cursor at 5kHz mark:
Here is measured HD of speaker with FH9 - the simple switch to vertical FETs with source/collector resistors and Vbe multiplier fixed the rising distortion problem seen above (same exact board):
Here is measured HD of speaker with Dx amp:
Here is measured HD of speaker with Inv JFET Circlophone amp:
Here is measured HD of speaker with FX8 Bimo mod - and in case you are wondering if something was amiss with the lateral FETs used in the FX8 5cm above that may be causing the rising distortion, the same exact later FETs from that amp were used in this amp:
Note that all the amps sound very good, excellent even. The Circlophone had the edge all along but now I think the FX8 Bimo may be close or better once I figure what is causing that mild -1.5dB depression in the response from 2kHz to 7kHz.
Attachments
Last edited:
Looks like the tweeter needs to be padded down some.
That rising response would probably make the speaker sound a little too bright for my taste.
Some people prefer that sound though.
The 1.5dB suck out is strange.
That rising response would probably make the speaker sound a little too bright for my taste.
Some people prefer that sound though.
The 1.5dB suck out is strange.
I have a full range ScanSpeak 10F/8424 in a 2-way FAST that may be better to use as a reference as it has very low HD and flat frequency response from 300Hz to 18kHz. I may repeat measurements with that and see what we get.
"So far, the FX8 Bimo mod appears to measure the best with my speakers in terms of lowest harmonic distortion (as measured with a microphone). The FX8 Bimo mod seems to have a slight 1.5dB depression in the response though from 2kHz to 5-7kHz as compared to the other amps which all track identically through that region. I wonder what can be causing this?"
Because THD lower at that frequency?
Because THD lower at that frequency?
I think it may be because I have not installed the output inductor so it acts like a padding resistor. I can test by bypassing the resistor.
I think that speaker would give a more accurate portrayal of the critical midrange and upper frequencies when comparing different amplifiers.
Only the 2R2 resistor making o/p impedance very high?, Best to compare with only a link right now OR after installing inductor.
reg
prasi
Last edited:
