[amb]

Quote:

Originally Posted by TzeYang /img/forum/go_quote.gif While on the other hand, the itch of building a Jisbos is no longer tolerable LOL. So i was wondering if i can just use the cheaper 2N5457/5460 with far worse ratings instead temporary just for experiments.

Patience is a virtue...

You're of course encouraged to experiment. I haven't done any work or simulations using the 2N5457/2N5460 for the JISBOS, nor have I built/tested a real circuit with them, thus at this time I can't offer any comparative performance data, or whether these JFETs are even appropriate to use.

 

[NelsonVandal]

Quote:

Originally Posted by TzeYang /img/forum/go_quote.gif I live in Malaysia. Shipping is a pain to pay for using my country's puny currency. I can't see myself spending any much more after splurging monies on the film caps and manly tools for case work recently. I might save up some more monies and buy around 10 pairs from you in near future i guess since i really do miss that kind of quality service you offer at your site. While on the other hand, the itch of building a Jisbos is no longer tolerable LOL. So i was wondering if i can just use the cheaper 2N5457/5460 with far worse ratings instead temporary just for experiments.

Use sockets for those JFET's like I did, and you can just plug in the better JFET's when you can afford them. You don't have to buy 10, 6 is enough if you buy matched ones. Even the unmatched I bought only differs a couple of mA's. You're in for a treat hearing these buffers.

 

[wui223]

TzeYang: i am also from Malaysia. maybe u can try Farnell for the Jfet. btw i live in KL, maybe can find those jfet at local stores. I am interested to try out this JISBOS buffer-amp also.

 

[Eokboy]

^Nope Farnell does not have them. Pasar only has 2SK170...

So, we should get a whole lot from amb, then share the shipping costs.

 

[wui223]

Eokboy: u tried every shop at JP? actually what jfet we need now for the JISBOS? if got people need it then bulk order from amb is fine.

 

[majkel]

This is true that virtual ground and ground channel are different points of the circuit sometimes but not in my case. I don't understand using extra buffering (with no feedback actually) just for headphones. This way channel feedbacks are referenced to a slightly different signal than the headphone output. It's weird and I don't understand why it is so good. Another disadvantage of having no feedback on the output ground is lowering the damping factor. I know it's not so critical for headphones but many models just sound better when there is no resistance on the output channels.
Regarding the noise - there is almost none on the highest volumes, it's significant on IEMs but you wouldn't listen to them so loud. Actually the noise is much lower than if I would use resistors of comparable values in the channel feedback. It's easily explainable - the input noise is not amplified while in the channels its amplitude multiplied by the voltage gain. I agree, it's better to use lower resistances when power consumption is not an issue, and increase the compensation capacitance accordingly.

 

[amb]

majkel, I don't know where you get the idea that the output ground channel amps that I mention has "no feedback"... They certainly do (in fact 100% feedback) and have extraordinarily low output impedances.

As for the separation of signal ground and output ground, it's to remove signal ground contamination by the high current return from the load. This has been discussed time and again in these forums as well as the official web pages of the respective amps, I recommend that you use the search function.

 

[majkel]

I dug search results for both "virtual ground" and "ground channel" phrases in the topic and found nothing valuable or even 1-post threads. I defer from walking thru all the topics containing any of the phrases in the posts. I guess I have to little life to finish reading.

I will argue that it's a 100% feedback when you have different reference points for signal input and signal output. For the latter you buffer the reference which is rather abnormal in standard designs where you struggle to have one solid ground, and you know the feedback voltages are referred to exactly the same ground. As long as you have efficient and fast (with high bandwidth in other words) virtual ground, you don't have to worry about "contamination by the high current return from the load" because it's equalized in the real time. I gave the simplest example how to make it work.
In PIMETA there is a different idea: Vs/2 is produced on the TLE2426 circuit. It is very slow and any battery voltage fluctuations due to alternating current load are not reflected on the virtual ground, so in this case "return current contamination" will occur, because of virtual ground being very slow, actually passive. So the idea was to isolate the slow virtual ground from dynamically changing currents. I showed the alternative. I hope you guys get the idea.

 

[NelsonVandal]

I'm back where I started, with a JISBOS as active ground. This time there's no offset drifting with long cables. No need for adding ferrites - yet

. I've added a 330 Ohm resistor between the output of TLE to the buffers input. I removed the 33 pF cap from output to ground. It seems stable both with battery and TREAD. Maybe it was bad solderwork before, but it sure looks messier now.

I've only listened to it briefly since I've spent hours on debugging it (I finally found out that a BC550 was bad). It sounds brighter. I think it's because OPA551 is so very dark sounding in comparison. It sounds very good, but it did even with the opamp as ground. If it keeps stable this will be the final rebuild. The stripboard is worn out, not a very good quality. I have to give it some more time to be sure of how it actually sounds. I'll show a new picture and report on how it sound tomorrow, have to go to bed now.

 

[NelsonVandal]

I'm to tired to do any real listening tests today. All work and no play...

 

[NelsonVandal]

There was a short from one of the batteries (the plastic covering is very thin, and there was a rip while squeezing them in) to the enclosure. I hate metal boxes. I wish there were better, more size effective and high quality, plastic boxes. I fried the TLE because of this, I think. So now I have to use resistors for voltage split/grounding. I have to get some new TLE's. I wonder if the TO92's are less sensitive.

I finally measured the current draw. It's "only" 120 mA when battery powered. The box gets nice and warm even when run on batteries

, and even warmer when wall-powered. Wall powered it's 160 mA + 100 mA for battery charging.

In spite of this unintended voltage devider, this amp sounds very good with JISBOS in all three channels. To my ears almost perfect. It's detailed yet warm and full, spacious yet down to earth and organic, and there's a lot of "oomph" in the bass after I increased the bias. It's better than any of my other amps. The only I have that can serve as some kind of reference are Mini3 (AD8397/OPA690), PPAS (LME497x0/LMH6321) and a LISAIII clone. The JISBOS is better.

After some more listening I think the quest for the pure and golden sound is over. This is so very good. I don't think I'll find a better portable device in the nearest future. The only thing missing is gain. I guess I'll have to wait for an ingenious simple all discrete amp or wait for the perfect opamp. Eventually there'll be better opamps. ADA4899-1 is very close.

 

[TzeYang]

ahhh, time to fabricate your own PCBs nelson

Do it for the portable buffer maniacs!

Even single layer is fine too. A fully incorporated Ground Plane would make it even sweeter.

 

[NelsonVandal]

Quote:

Originally Posted by TzeYang /img/forum/go_quote.gif ahhh, time to fabricate your own PCBs nelson Do it for the portable buffer maniacs! Even single layer is fine too. A fully incorporated Ground Plane would make it even sweeter.

Several reasons not to:
1. There'll soon be "Amb quality" JISBOS PCB's. No need for anything else. I think they could be placed inside a fairly small enclosure.
2. It's not safe to squeeze batteries and electronic components like this inside a small metal box. Not even I dare to charge this overnight - yet. I wouldn't want anybody's house to burn down

.
3. There seem to be no portable buffer maniacs except me. People don't know what they're missing. Using amps with gain on IEM's and low Z phones is just silly. PPL was about to do a portable unity gain amp especially for IEM users, and there were no interest at all.
4. The JISBOS is very, I mean very, easy to do on a stripboard. They're easier than the diamond buffers. They're very symmetrical, no cable crossing, no need for cables at all for the internal wiring. Cables are only needed for connecting the rails and in/out. Using stripboards with a ground layer makes things easier.

TzeYang, I know you're doing a JISBOS amp as well. I hope to hear from you when you've finished it, and compared it to your PPA.

 

[kvant]

Quote:

Originally Posted by NelsonVandal /img/forum/go_quote.gif The only thing missing is gain. I guess I'll have to wait for an ingenious simple all discrete amp or wait for the perfect opamp.

Maybe there is a way how to squeeze some gain from JISBOS. It seems to me that Hiraga's "Le Monstre" have a lot in common. "Le Monstre" has a gain >1. Perhaps one could learn something from that design.

 

[NelsonVandal]

Cool. I've only heard of Hiraga, never heard any of the amps. He was very highly regarded in the 80's when I worked in a HiFi shop. I'm gonna use Google translate and try to read this, but I think unity gain = gain 1.