FX8 50mm Stereo - listening impressions and measurements
I connected this to my usual 35v 250VA Dx PSU, first used 10R safety resistors to debug. Upon startup, I get about 600mA (6 volts) across the 10R resistors. Quickly turning trimpot clockwise (like shutting off a spigot), the bias currrent drops. I set at 175mA - stable after 5 min of initial warmup. Output DC offset is 0.9mV to 1.2mV - matching the Hfe on the BC546's pays off here.
I am using a different speaker to test the sound quality - my reference active monitors using ScanSpeak 10F/8424 for mid/tweets and RS225-8 for bass duties. Xo and EQ is provided by miniDSP 2x4 running a Harsch XO at 500Hz in FAST configuration for quasi transient perfect setup. In this way two amps are required, one for woofer and one for mid/tweet. Mic is Cross Specrtrum Labs calibrated UMIK-1. Measurements made at 0.5m distance to avoid floor bounce.
More info on the speaker used here:
http://www.diyaudio.com/forums/full-range/273524-10f-8424-rs225-8-fast-ref-monitor.html
So in summary, it sounds FANTASTIC! 😀
The cheap Chinese lateral FETs from Aliexpress are the real deal. Anyone who assumes that cheap Chinese means fake and no-good needs to re-think that assumption. Perhaps the trick is using all film caps and the 10uF MLCC ceramic for the input? I don't know but this amp measures better than my previous best amp - the JFET Circlophone. Slightly lower HD, and deeper bass extension. The SS 10F/8424 is a very low distortion driver capable of picking up on these subtleties.
So first, here is a photo of the working amp. I am using four 220R resistors in parallel to drop the cooling fan speed to inaudible but cooling is still very effective (barely above ambient).
Here is the comparison of the FX8 50mm vs the JFET input Circlophone with the 10F/RS225 FAST speaker showing the cross over response. FX8 is the light blue line and note the deeper bass extension. You will notice that the combined response is a bit lower than the woofer only - that is normal and evidence that the Harsch XO is setup properly:
Here is the HD for the JFET input Circlophone (cursor is at 1kHz):
Here is the HD for the FX8 (cursor is at 1kHz), note slightly lower HD here, and the peak near 120Hz might be ambient noise:
Finally, attached is a quick sound clip of the FX8 recorded with a Zoom H4 and converted to mp3. Change file extension from .asc to .mp3 to listen. The clarity of the saxophone is quite astonishing, as well as the textures.
I connected this to my usual 35v 250VA Dx PSU, first used 10R safety resistors to debug. Upon startup, I get about 600mA (6 volts) across the 10R resistors. Quickly turning trimpot clockwise (like shutting off a spigot), the bias currrent drops. I set at 175mA - stable after 5 min of initial warmup. Output DC offset is 0.9mV to 1.2mV - matching the Hfe on the BC546's pays off here.
I am using a different speaker to test the sound quality - my reference active monitors using ScanSpeak 10F/8424 for mid/tweets and RS225-8 for bass duties. Xo and EQ is provided by miniDSP 2x4 running a Harsch XO at 500Hz in FAST configuration for quasi transient perfect setup. In this way two amps are required, one for woofer and one for mid/tweet. Mic is Cross Specrtrum Labs calibrated UMIK-1. Measurements made at 0.5m distance to avoid floor bounce.
More info on the speaker used here:
http://www.diyaudio.com/forums/full-range/273524-10f-8424-rs225-8-fast-ref-monitor.html
So in summary, it sounds FANTASTIC! 😀
The cheap Chinese lateral FETs from Aliexpress are the real deal. Anyone who assumes that cheap Chinese means fake and no-good needs to re-think that assumption. Perhaps the trick is using all film caps and the 10uF MLCC ceramic for the input? I don't know but this amp measures better than my previous best amp - the JFET Circlophone. Slightly lower HD, and deeper bass extension. The SS 10F/8424 is a very low distortion driver capable of picking up on these subtleties.
So first, here is a photo of the working amp. I am using four 220R resistors in parallel to drop the cooling fan speed to inaudible but cooling is still very effective (barely above ambient).
Here is the comparison of the FX8 50mm vs the JFET input Circlophone with the 10F/RS225 FAST speaker showing the cross over response. FX8 is the light blue line and note the deeper bass extension. You will notice that the combined response is a bit lower than the woofer only - that is normal and evidence that the Harsch XO is setup properly:
Here is the HD for the JFET input Circlophone (cursor is at 1kHz):
Here is the HD for the FX8 (cursor is at 1kHz), note slightly lower HD here, and the peak near 120Hz might be ambient noise:
Finally, attached is a quick sound clip of the FX8 recorded with a Zoom H4 and converted to mp3. Change file extension from .asc to .mp3 to listen. The clarity of the saxophone is quite astonishing, as well as the textures.
Attachments
Last edited:
Zoky2
Can you provide details of the SMPS?
Is it a diy module?
Sent from my HUAWEI MT7-L09 using Tapatalk
"The cheap Chinese lateral FETs from Aliexpress are the real deal. Anyone who assumes that cheap Chinese means fake and no-good needs to re-think that assumption."
This certainly isn't true all of the time. I have purchased a good number of transistors and op-amps from Chinese sellers over the years.
Some of the transistors were checked on a curve tracer by another member and verified to be fake.
When it comes to getting real Toshiba JFETs and MOSFETs from Asian sellers, you can pretty much count on getting fakes every time.
They also counterfeit expensive op-amps like the OPA627 and AD797 as well as cheaper ones like the OPA2134.
As the ole' saying goes, "If the price seems to be too good to be true then it's probably a fake".
On your FX8 amps, exactly what component changes have you made to the original circuit?
Are you using BD139 and 140s?
I've almost got my boards done and would like to install the right components the first time.🙂
Thanks...
This certainly isn't true all of the time. I have purchased a good number of transistors and op-amps from Chinese sellers over the years.
Some of the transistors were checked on a curve tracer by another member and verified to be fake.
When it comes to getting real Toshiba JFETs and MOSFETs from Asian sellers, you can pretty much count on getting fakes every time.
They also counterfeit expensive op-amps like the OPA627 and AD797 as well as cheaper ones like the OPA2134.
As the ole' saying goes, "If the price seems to be too good to be true then it's probably a fake".
On your FX8 amps, exactly what component changes have you made to the original circuit?
Are you using BD139 and 140s?
I've almost got my boards done and would like to install the right components the first time.🙂
Thanks...
Well the HD on the Chinese lateral FETs is as low if not lower than the Renesas from a verified source in the US. They were less expensive but not as cheap as a vertical FET. $5.40/pair free shipping. I am going to buy more - if anyone else wants to try here is link.
http://m.aliexpress.com/item/32480898835.html
One thing to note is that the metal tab is solid copper and milled nicely. The pins are smooth and tinned with solder (not crisply stamp cut and un-tinned), the epoxy is nicely formed and clean. By all accounts - it physically looks like a real part. Electrically it is doing what it was supposed to.
On this last build of the 50mm FX8 I am using 100uF 63v power supply rail caps vs 47uF. I am using 330uF 25v OSCONs instead of 1000uF 16v. I am using BD139/140 marked NXP from China. BC546C's and MPSA92 also from China bought in a bag of 50. They may all be fakes and they may not be, as when I buy parts from Digikey, the conformance certificate shows about 70% of items originate from China. So many of the parts probably come from same OEM factory. I only know that with this build the parts fake or real are performing just fine for a lot less than equivalent Mouser/Digikey/Newark sourced items.
http://m.aliexpress.com/item/32480898835.html
One thing to note is that the metal tab is solid copper and milled nicely. The pins are smooth and tinned with solder (not crisply stamp cut and un-tinned), the epoxy is nicely formed and clean. By all accounts - it physically looks like a real part. Electrically it is doing what it was supposed to.
On this last build of the 50mm FX8 I am using 100uF 63v power supply rail caps vs 47uF. I am using 330uF 25v OSCONs instead of 1000uF 16v. I am using BD139/140 marked NXP from China. BC546C's and MPSA92 also from China bought in a bag of 50. They may all be fakes and they may not be, as when I buy parts from Digikey, the conformance certificate shows about 70% of items originate from China. So many of the parts probably come from same OEM factory. I only know that with this build the parts fake or real are performing just fine for a lot less than equivalent Mouser/Digikey/Newark sourced items.
Last edited:
All
Is 5.4$ for a pair of 2sj and 2sk? Site says per piece. Just want to confirm.
Hi X,
I'm not familiar with the OSCON caps. Are these what you're using?
25SEPF330M Panasonic | Mouser
If I use these "Gold" Nichicons for the rail caps, do you think the sound will be even better?🙄 Just kidding.😀
What did you use for the 10uF input cap?
That is the correct cap - SEPF 330uF 25v. These are very low ESR caps and a favorite that I used for power rail caps on class D amp. It helps improve transients and large bass excursions.
Better rail caps probably would help. I am using cheap Chinese ones 50 per bag for $2.87
The 10uF input cap is a no name MLCC rated at 50v. Probably 50 for $2.
Here is a closeup of the boards so you can see the components better.
Attachments
Last edited:
Hi ammel68.
if i could give an advise to you, i have in last finished Amp done a taste with Cornell DME
pp- not high prized, but really listenable.
Also you can parallel a 0,01µ- 10n orange drop 715P- i like that since i tested in decoupling in tube amps.
The Panasonic now was a Sanyo in the past, all sold.
The ultra low ESR Cap you can see in use decoupling Katode in tube amps also.
For that reason i mean there must be anywhere a signal/sound influence- not all voodoo.
More i don't could say, about not knowing deeper experts knowledge.
Regards Bangla.
P.S. Sorry for typo error!
The nice thing about the layout of the FX8 Bimo board is the roomy provision for large input caps. I could put an axial polypropylene cylinder 10uF 450v cap if I wanted to. With this 50mm board, compactness was the main driver. So choices of caps that fit are limited. Certainly a small 0.1uF film bypass cap could be added and probably might help. I may try that next. I think a 4.7uF 63v MKP Wima will fit in that spot though. The thru hole there is extremely small so caps with thick pins won't work.
Edit: I found link to the 10uF cap I am using.
http://m.aliexpress.com/item/32455168081.html
Edit: I found link to the 10uF cap I am using.
http://m.aliexpress.com/item/32455168081.html
Last edited:
wow, fh9 dethrone circlophone? more bass? :S
makes you wonder how much better apex's more advanced designs are. i was initially going for sr150 but was put off by the complexity and the comment that it sounded like a 'studio amp', comparable to symasym.
i'm well on my way to etching fh9 pcbs, btw.
makes you wonder how much better apex's more advanced designs are. i was initially going for sr150 but was put off by the complexity and the comment that it sounded like a 'studio amp', comparable to symasym.
i'm well on my way to etching fh9 pcbs, btw.
Can higher power versions of 2SK1058 and 2SJ162 be used on FX8 for more power?
For example 2SK1529 and 2SJ200?
For example 2SK1529 and 2SJ200?
Thank you so much for posting the pdfs Sir....
I am going to etch the pcb....
People like you are real inspiration to try out new amps....
I am going to etch the pcb....
People like you are real inspiration to try out new amps....
You can get ECW20N20 and ECW20P20 double die 250W Lateral mosfet from www.profusionplc.com but they are not cheap.
I asked because I saw this simple 10W class A amp by Widowmaker, posted by Lineup where he says these are lat FETs.
http://www.diyaudio.com/forums/soli...e-end-jfet-input-lateral-n-mosfet-output.html
Also, this thread seems to imply they are laterals.
http://www.diyaudio.com/forums/soli...l-mosfets-class-ab-overview-p-spice-urls.html
It's such an important distinction yet nowhere to be found in the factory spec sheets. Why is that?
Second channel builded and tested with stabilized PSU. On my cheap speaker I can not hear the difference between linear PSU and SMPS. Will test more next week when i have more time.
About SMPS ,yes its a diy proyect from elitesecurity.org. Most parts are from old computer PSU,the transformer is revinded to nedeed voltage.The language on elitesecurity is not english so you may have problems to undestand that site.
regards
zoky2
About SMPS ,yes its a diy proyect from elitesecurity.org. Most parts are from old computer PSU,the transformer is revinded to nedeed voltage.The language on elitesecurity is not english so you may have problems to undestand that site.
regards
zoky2
