DIAMOND DIFFERENTIAL INPUT POWER AMPLIFIER EBay kit

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
No one says on eBay what they are selling is a clone of a commercial product, present or past.

Whatever the word they use (resemblance, reference, inspired on) it's only a matter of semantics. If the people who should be buying the kit do not read on what existing product is it based, they wouldn't buy it. So that characteristic has to be quite clear.

At the beginning of this thread it was discussed why was it called Diamond differential, and some comparisons were made with Sansui amps. But they were not. I still do not know the reason for the name.

Actually, it's probably a legal requirement, since making a duplicate of a commercial amplifier would be a violation of copyright.

Here's a thread about the Sansui design:
http://www.diyaudio.com/forums/soli...nt-mirror-loads-self-biased-vas-cascodes.html
 
You're certainly right about the eBay legal requirements.

That Sansui does resemble the E305, but not the E305V.

Yeah, eBay enforces the copyrights of other companies. I got stung by that once when I tried to sell a Rosetta Stone language learning software there.

I can't find much information on a "diamond differential amplifier". I've found plenty on the diamond buffer, which appears to be two pairs of opposite polarity cascaded emitter followers, but I don't pretend to understand the why and wherefore of it.
 
In the other thread about the diamond buffer, a SPICE software model called Tina was recommended. So I downloaded it from ti.com. It is easy to use. I got adventurous and built a diamond buffer with CCS loads. It checks out at DC but the Fourier analysis says it has 495% distortion.... hmmm.... something not right there!
 
I have checked all transitors, driver and power. Power I'm using NJW1302g/3281g. These were tested on a few Ic currents from 100uA up to 20mA.

In the original circuit there was no bd139 but sc5171 but I have many bds from Mouser.

cheers,

Try simulating the circuit in Tina or another SPICE program.

I just finished a diamond buffer that has extremely low distortion.
 
It was simulated using LT spice. Results were pretty good but VAS was susceptible to oscillations if PS was of poor quality. From what I experience it does not look like oscillations as huge oscillations would heat up transistors very quickly as well as zoebel resistor.

cheers,
 
It was simulated using LT spice. Results were pretty good but VAS was susceptible to oscillations if PS was of poor quality. From what I experience it does not look like oscillations as huge oscillations would heat up transistors very quickly as well as zoebel resistor.

cheers,

Is that the schematic in the first post? If so, why was W2 deleted? Isn't it for DC offset? Maybe W1 is for DC offset. W3 appears to be for setting the bias in the outputs. I admit I don't understand the second stage, but basically, I think that one way you can improve the PSRR is by using CCD's in certain places instead of resistors.

I don't even see a VAS in this schematic. I see a super fancy differential amplifier at the input, providing balanced signal output to a super fancy emitter follower, which in turn drives the outputs. All the voltage gain seems to be in the first stage. I can't for the life of me figure out how gain can be obtained from Q9 and Q10, for example.
 
Last edited:
I have kept W2. LT spice simulations gave best results with 220 ohm on both sides, that is R21=220 and R16 || W2 plus r28 =220 ohm so I set them to be 220. In reality I cut the connection of nfb to the centre between R22 and R23. Drivers are floated. That is theoretically a better solution.

I haven't had time to do the intended/suggested few mods to the circuit and try it again.

cheers,
 
I think you should first restore the original connection to R22/R23, and see what happens.

I also think there might be a mistake on where they are joining the output path, as the NFB resistor is in between.

It looks as if the floating resistor that might have been on the Accuphase was split as a last moment thought. AFAIK the NFB should be after that junction or completely separate.
 
Diamond v1.2 JIM AUDIO

Hi everyone,

I bought a kit Diamond Differential amp version 1.2 Store JIM AUDIO. I rode carefully made ​​adjustments without powers OnSemi NJW0281G and NJW0302G. The trimpot W2 does not change anything, can turn thousand rounds for each side. After soldered the powers and turned 48.7 VDC: but nothing, no audio comes out!!!

Thanks, I'm newbie, I have difficulties. Any suggestions please ? :crazy:
 

Attachments

  • Diam1.2-jim-audio2.jpg
    Diam1.2-jim-audio2.jpg
    187 KB · Views: 594
  • Diam1.2-jim-audio1.jpg
    Diam1.2-jim-audio1.jpg
    227.3 KB · Views: 568
Set W2 to such a value that together with R16 and R23 (330ohm in series with W2) this part of the circiut becomes 220 ohm. Alternatively, remove W2 and change R16 to 220 ohm. Actually the circuit should react to changes in W2. Mine did.

Now, overall it looks that the circuit is wrong. I mean it seems to work when simulated but it does not work in reality. One of the problems is the bias circuit but there are possibly a few other problems. Carlmart worked on redesigning the circuit with a friend of his but I do not think he has built his modified amp yet. It think he is still playing with it on his PC.

I do not know if anyone here managed to make this amp operational. I'm still working on it. Modified versions work on a simulator but not necessarily in practice. The original worked in LT spice although it was very susceptible to oscillations and quality of PS.

Write to Jim's Audio and ask them for assistance. I'm afraid this amp in the current version is a dud. It needs some parts of its circuit redesigned.

cheers,
 
Thanks for the effort to help. Stanton Tin (JIM AUDIO) has undertaken extraordinary effort, with many e-mails and photos, examining the position of the parts and checking the voltages that I measured here at my house. Tin explained that w2 does not work because the boards do not have anything in the output.

I noticed a curious question, can you clarify: two different schematic, the first driver Q0 is D669 and resistors RF1 (22k) and RF2 (1k) have physical connectivity with R35 and R36 (point C in its recommendations for adjustments)

The second schematic the driver transistor Q0 is the C5171, but the joints of the R35 and R36 with both RF1 and RF2 does not exist: the transistors Q4 and Q6 are connected in RF1 and RF2 but are NOT soldiers in two resistors 33R / 0.5 watt (R35 and R36)?

It can not be the problem because the gates of Q4 and Q6 should not be connected to R35 and R36?

Other suggestions are welcome. Again, thanks to the best participants of this forum, special janusz, MOER and Stanton Tin (JIM AUDIO).
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.