Hugh,
Thanks for the confirmation. The other channel ended up having -2mV offset and about 2.3 mA bias for the input stage LTP.
I also tested this channel but went a little further and did 4 ohm testing. Mind you my bench supply was set at +/- 24VDC and my source resistors are 0.25 ohms each which sets the current at ~ 1.35A.
FRA/Bode Plot analysis revealed 10Hz-200khz (-3dB) bandwidth with an 8Vp-p input (1 watt/8 ohms). Overall gain is ~19.5dB, exactly what I wanted. Phase was flat to nearly 20khz (-7 degrees). Square waves at various frequencies looked just fine for a SEPP design.
So with the aforementioned test conditions, It clipped at 47V p-p or 33 watts RMS with an 8 ohm load and right at 23-24V p-p or 17 watts RMS for a 4 ohm load. This time around however, I watched to see If that cool RED led “clip” indicator would come on and it DID! It was exactly at 47V p-p for the 8 ohm load but interestingly enough for the 4 ohm load, it was more like 35V p-p. At this level, the trough of the sine wave had severely clipped (it actually starts at 23.5V p-p) and the peak of the sine wave was just starting to clip. Fascinating to see, and interesting given the topology.
Next, I need to remove the ALPHA 20 boards in my NPXP and install the Alpha Nirvana for a listening session. If this is better than the ALPHA 20 design then I might consider installing the Alpha Nirvana in its own chassis and +/- 28V supply voltage, plus slightly smaller source resistors. In addition, a very tiny hole up front just for that cool clip indicator 😉.
It would be cool if we can configure it with a bicolor LED that goes from green to red when clipping occurs.
Best,
Anand.
Thanks for the confirmation. The other channel ended up having -2mV offset and about 2.3 mA bias for the input stage LTP.
I also tested this channel but went a little further and did 4 ohm testing. Mind you my bench supply was set at +/- 24VDC and my source resistors are 0.25 ohms each which sets the current at ~ 1.35A.
FRA/Bode Plot analysis revealed 10Hz-200khz (-3dB) bandwidth with an 8Vp-p input (1 watt/8 ohms). Overall gain is ~19.5dB, exactly what I wanted. Phase was flat to nearly 20khz (-7 degrees). Square waves at various frequencies looked just fine for a SEPP design.
So with the aforementioned test conditions, It clipped at 47V p-p or 33 watts RMS with an 8 ohm load and right at 23-24V p-p or 17 watts RMS for a 4 ohm load. This time around however, I watched to see If that cool RED led “clip” indicator would come on and it DID! It was exactly at 47V p-p for the 8 ohm load but interestingly enough for the 4 ohm load, it was more like 35V p-p. At this level, the trough of the sine wave had severely clipped (it actually starts at 23.5V p-p) and the peak of the sine wave was just starting to clip. Fascinating to see, and interesting given the topology.
Next, I need to remove the ALPHA 20 boards in my NPXP and install the Alpha Nirvana for a listening session. If this is better than the ALPHA 20 design then I might consider installing the Alpha Nirvana in its own chassis and +/- 28V supply voltage, plus slightly smaller source resistors. In addition, a very tiny hole up front just for that cool clip indicator 😉.
It would be cool if we can configure it with a bicolor LED that goes from green to red when clipping occurs.
Best,
Anand.
Hi Anand,
We have several amps that use off board snubbers with flying leads:
Alpha Nirvana, Alpha BBB, FH9HVX, and future amps. We like the ease of separating the output mounting process and the PCB mounting process. I believe the FH9HVX being a Class AB has even the same board footprint and flying lead pin outs so you could theoretically just swap boards, and PSU voltages.
FH9HVX on my griddle amp:
We have several amps that use off board snubbers with flying leads:
Alpha Nirvana, Alpha BBB, FH9HVX, and future amps. We like the ease of separating the output mounting process and the PCB mounting process. I believe the FH9HVX being a Class AB has even the same board footprint and flying lead pin outs so you could theoretically just swap boards, and PSU voltages.
FH9HVX on my griddle amp:
Last edited:
X,
I stand corrected! Thanks for clarifying and hope to see more designs from your collaborations.
Best,
Anand.
I stand corrected! Thanks for clarifying and hope to see more designs from your collaborations.
Best,
Anand.
Hi Hugh,
I'm a bit confused! 🙁
Looking at the graph which Rf posted in post #2108 ... the line is already drooping at 1ma. So if we want to stay in the 'linear zone' ... why wasn't the current thru R111. R112 & R113 set at 2ma - or even less - rather than 2.6ma?
Thanks,
Andy
You are confused Andy, there are so many schemats on this amp I am too.
Let's go back to the original on post #1, V2 of the LTSpice schemat drawn up by X and me.
You are right. It uses the LTP arms of 220R and 47R. In that circuit, the LTP stage current is 2.6mA.
My mistake........
Hugh
Let's go back to the original on post #1, V2 of the LTSpice schemat drawn up by X and me.
You are right. It uses the LTP arms of 220R and 47R. In that circuit, the LTP stage current is 2.6mA.
My mistake........
Hugh
Yes, LTP stage current is 2.6ma.
So why not 2ma or less ... given the graph in post #2108?
Andy
Andy,
It would not make much difference!!
Using bipolars in LTPs beta is not too important; it merely affects the base current.
The most significant in this role is the transconductance; voltage change at base for current change at collector.
HD
It would not make much difference!!
Using bipolars in LTPs beta is not too important; it merely affects the base current.
The most significant in this role is the transconductance; voltage change at base for current change at collector.
HD
Looking good! Those power resistors for the outputs are new to me..?
I like to 'flatten' the output devices on class A amps to improve heat transfer, the surface can be surprisingly uneven. I think a flat file is good enough, but there are more sophisticated ways of course..
I like to 'flatten' the output devices on class A amps to improve heat transfer, the surface can be surprisingly uneven. I think a flat file is good enough, but there are more sophisticated ways of course..
RF,
Yes, I use some sandpaper on the rear of the MOSFETS prior to thermal grease + Al Oxide spacers. I don’t use thermal insulators like Keratherm on Class A amps that have high bias settings typically. On Class AB and low bias Class A, sure.
The source resistors here are Riedon brand FPR type (CuMgNO3). Expensive though. I have the KOA Speer BPR stock ones as well, but the Riedon did fine in bench testing and temps were ~ 32 deg C.
Best, Anand.
Yes, I use some sandpaper on the rear of the MOSFETS prior to thermal grease + Al Oxide spacers. I don’t use thermal insulators like Keratherm on Class A amps that have high bias settings typically. On Class AB and low bias Class A, sure.
The source resistors here are Riedon brand FPR type (CuMgNO3). Expensive though. I have the KOA Speer BPR stock ones as well, but the Riedon did fine in bench testing and temps were ~ 32 deg C.
Best, Anand.
Super work, Anand! Those boards look great.
I just ordered some clone BPRs from AliExpress. Will test them out to make sure they don’t contribute high distortion. But they were plentiful like tap water. 200 for $20.
I will crack one open to make sure it is non inductive bulk metal and not a wire wound. Then connect 8 x 0.47R up as 4ohm dummy load for an amp distortion test using one of the lowest distortion resistors out there - EBG UXP300.
I just ordered some clone BPRs from AliExpress. Will test them out to make sure they don’t contribute high distortion. But they were plentiful like tap water. 200 for $20.
I will crack one open to make sure it is non inductive bulk metal and not a wire wound. Then connect 8 x 0.47R up as 4ohm dummy load for an amp distortion test using one of the lowest distortion resistors out there - EBG UXP300.
Anand,
Brilliant work, and superb photography. It really brings it out when you use skillful pictures!
And very nicely made modules, congratulations!
HD
Brilliant work, and superb photography. It really brings it out when you use skillful pictures!
And very nicely made modules, congratulations!
HD
The fake BPRs work well as CRC resistors on PSU’s but I would stay away from using them in amp output stages. More here:
https://www.diyaudio.com/community/...bench-tonight-obt.372508/page-28#post-6918791
https://www.diyaudio.com/community/...bench-tonight-obt.372508/page-28#post-6918791
+1
Anand, just another question - I see on your board that you jumpered R136 (4R7) and not R133, R135 (100R carbon) as I understand it should be?
Nice input caps!
2c,
Yes, I jumpered R136 as the resistor size I had ordered was subsituted by Mouser to a larger 1 watt size. R136 is there for minimizing the chance of oscillation, not absolutely necessary but recommended. If you look back in the thread, Do (pinnochio) did the same. No issues thus far in my testing on my scope.
R133, R135 are gate stopper resistors and are optional if you have already populated the gate stoppers on the individual MOSFET snubber boards. The gatestoppers on the snubber boards must be populated however (they are not optional). The audible differences between 100 ohms vs 200 ohms as gatestoppers is probably negligible. They are also there to minimize oscillation risks.
Best,
Anand.
Ok great, so now I can relax about your build and mine to come. You obviously understand the role and workings of each component, which I don't. I paint by numbers still.
Looking very forward to your NPXP listening impressions.
Looking very forward to your NPXP listening impressions.
2c,
Thanks. Tonight or when I have time, I am going to post the schematic and systematically itemize what parts I used where and why. I have been getting some questions regarding this. This is all somewhat subjective but I am quite sure that none of my substitutions will detract from the overall measurements and stability of the amplifier. Thanks for reading!
Best,
Anand.
Thanks. Tonight or when I have time, I am going to post the schematic and systematically itemize what parts I used where and why. I have been getting some questions regarding this. This is all somewhat subjective but I am quite sure that none of my substitutions will detract from the overall measurements and stability of the amplifier. Thanks for reading!
Best,
Anand.
I've just finally had time to listen to the amplifier! Work has got in the way and I was too tired at night to listen to anything aside the inside of my eyelids... 🙂
Very quickly, for early observations... Superb soundstage, very good 3d positioning of sounds/instruments, also excellent vocals, and details un the upper range. Some details, from some albums/songs, usually a bit more hidden/buried, I've heard coming out to the light in a very pleasant fashion, not getting in the way and not getting at all in the analytical, just very enjoyable!
Tonight I'll have some time to listen a little more.
Do
Very quickly, for early observations... Superb soundstage, very good 3d positioning of sounds/instruments, also excellent vocals, and details un the upper range. Some details, from some albums/songs, usually a bit more hidden/buried, I've heard coming out to the light in a very pleasant fashion, not getting in the way and not getting at all in the analytical, just very enjoyable!
Tonight I'll have some time to listen a little more.
Do
Here are some cliff notes as I built this board design. We'll start from the left and then move to the right of the above schematic. Even though it appears to be complex, once you break it into sub circuits, it's much easier to understand and actually a simple circuit. Remember there are only 6 standard active transistors in this circuit. My substitutions were specific with regards to the 'AC' music signal.
R101 and R102 are part of the input RFI filtering stage of the circuit and are Takman REY metal film.
C101-C105 are actually all the same capacitor. In the schematic, X gives you options for various sizes of capacitors and accommodates electrolytics, film, etc...This is one of the most important caps in the circuit. As such, if you can splurge...splurge here! I used a combination of Vcap ODAM & Vcap CuTF. An overall value of 5uf to 10uf would be best to get extended low frequency response.
For V111, and V112, I had a long reel of KSA992's and found a quad match (for Hfe) which I used.
For R112, I substituted a 5K CMF55 resistor and then used a pot for R113 to dial in the requisite bias for the input LTP stage. I measured the bias by placing my voltmeter across R111 (3k3 CMF55 resistor), while carefully watching what my other voltmeter on the output of the amplifier was with regards to DC offset. Ideally, you want something with reasonably low DC offset (less than 20mV is fine), and bias to be around 2mA. Use V=IR to calculate what voltage you would need across R113 to get that bias. Eventually you'll want to replace the pot with a resistor value equivalent.
For C114/C115 which is again the same capacitor, I used a 15pf silver mica cap. You can also use an SMD NPO cap if you wish.
For C111, I used a Nichicon Z series capacitor. This is a bypass cap for the input LTP stage, it should be of good quality to filter out ripple.
In the center of the schematic are some very important passive parts for this circuit. They are parts of various feedback networks and definitely contribute to the sound of the amplifier. These are R124, R126, C124, R123, C121, and C1250. Again for C124, C121, and C1250, X has provided alternative parts substitutions to accommodate various budgets and desires.
R124 and R126 are directly involved in setting the overall gain of the amplifier. I like amps that are around 15 dB to about 22 dB of overall gain. The reason has to do with the overall sensitivity of my speakers (96dB), the fact that I have a high sensitivity distributed multisub system, and that my source (DAC) output voltage is 5V RMS at 0 dBfs. So I have plenty of drive. Most dacs measure best when they are playing at nearly full tilt, as such, there isn't a need for a preamp for DAC only systems. It isn't recommended either especially if you are fanatical about maximizing signal to noise ratios. This also has to do with your listening position, and your music selections. In my experience, I typically have my Roon volume control at around -5 to -10 dB (relative to 0 db which is max volume level) when I am jamming at 85 to 100dB at the listening position. So choose the gain that is most appropriate for your setup and listening preferences. Of course folks do like various amounts of "flavor" in their systems which is where active preamps whether tube or solid state have their roles.
For R126, I kept it at 22k ohms. For R124, I set it to 2.2k ohms. The combination of the two gives me ~ 19.5dB gain. I chose Audionote tantalum film resistors here. They are notoriously expensive though. For C1240 I chose an Audionote Kasei capacitor. A minimum of 100uf is recommended here.
Next, for R123, C121 and C1250, I used a 330k CMF55 resistor, a 4700pf WIMA cap and a 10pf silver mica cap. This is also an area of experimentation which I believe member Rallyfinnen has been playing with. Look to his posts for more info.
As discussed previously in this thread, R136 is optional but recommended. I didn't have any small value resistors in stock, so I just shorted it. It's a nicety to minimize any potential for oscillation, so if I were to find a small value resistor I will reconsider populating it, although it's not in Hugh's original schematic.
For C131 and C132, I stuck with Elna Silmic 220uf values. You want to use a good quality electrolytic here as well.
R133 and R135 are gatestopper resistors and are optional if you have already populated R101 on the MOSFET snubber boards. I used 100 ohm Takman REY metal film for these resistors. It's best to always have a gatestopper resistor as close to the MOSFET gate pin to minimize oscillation risks.
R141 and R142 have been discussed previously in this thread (first 10-15 pages where X did a multitude of measurements and experiments). These are the source resistors for the output stage MOSFETS. They do have an influence on overall sound and distortion. In these positions I am still experimenting but I have chosen Riedon FPR2 series which have a copper magnanin foil that is seen in some expensive resistors of the Vishay VPG division. Values here can be 0.18 ohms to 0.25 ohms. Of course at lower values, bias current is higher and vice versa. At 0.25 ohms, my bias current is ~ 1.35A. I plan on trying 0.22 ohms in the future but the only value that was in stock at the time of purchase was 0.25 ohms.
For V141 and V142, I used FQA Mosfets of course!
C151 and C152 are local supply bypass capacitors for the +/- voltage rails, I chose 100uf Nichicon Z series I had on hand.
Hope this helps, Anand.
Last edited:
