PB2 said:
Hi Pete,
If you had an email button I would have sent you a note.
I appreciate all of the help I have gotten on the many projects I have done since coming here. It should not be overlooked that this is a DIY site where we all are supposed to design, build and learn. I don't think I'm overstepping my bounds here by saying that I have given back as much or more than I have taken from here.
Now I tried to put a bit of a humorous spin on the situation here, but I can see that it went right over your head. As usual.
We are all different, I know. My reaction to someone wanting to take one of my projects and change it or improve it would be one of interest and enthusiasm. I wouldn't feel threatened by this. I wouldn't say: "do what you want, but don't refer to it in the same breath as mine". I'd feel flattered and vindicated that something that I did here made some difference.
KLe said:
I will have to take the plunge and board this up so that I can test it. The simulator does a good job of pointing in the right direction but the truth comes out in the real thing. I think the THD figures are close to accurate, but I don't put a lot of faith in the distribution on the harmonics. I think the simulator is just "guessing" there.
Still, low is low and better than high, right?
KLe said:
I am not sure about the name BAIN though
That popped into my head while I tried to think up a name for the sim file. Fits with the theme I think.
Anyway, tis too late - I've already busted a champagne bottle on her bow...
You have a way of dropping insults, it went over my head?
It is said that there is always some truth in humor, and it is your underhanded way of continuing to annoy him because you didn't like his reaction. There are much bigger things in life - get over your dislike of his response. You could simply respect his wishes and get out of his thread quietly without drama. Which I will now do with yours - not worth the time.
Enjoy your hacking and patchwork, but do the forum a favor and stop deliberately annoying people who have made significant positive contributions.
You say you appreciate the help that you've gotten, talk is cheap, I judge people by their actions.
Please don't email me.
Pete B.
It is said that there is always some truth in humor, and it is your underhanded way of continuing to annoy him because you didn't like his reaction. There are much bigger things in life - get over your dislike of his response. You could simply respect his wishes and get out of his thread quietly without drama. Which I will now do with yours - not worth the time.
Enjoy your hacking and patchwork, but do the forum a favor and stop deliberately annoying people who have made significant positive contributions.
You say you appreciate the help that you've gotten, talk is cheap, I judge people by their actions.
Please don't email me.
Pete B.
MJL21193 said:
MJL21193 said:
MJL21193.
As your nice amplifier is close to AD797 input and 2nd stage
you may some time when you wish,
try to use the AD797 second stage idea.
See attachment. Also same diagram can be found in AD797 datasheet.
This is a jcx posted variant of AD797.The LED in second stage is not in the original.
The idea is to use a floating mirror - gives at all output voltages same Vce across Q8 and Q9.
This is one of the secrets of AD797 exceptional performance.Lineup
regardsAttachments
This is what "got me on the run" .. to explore how
your amp and mine handles mild and gross clipping.
The abomination does the best, it has less gain , just asymmmetric as noted above (just a little sticking)..livable .Mine sucked when overdriven.(both asym. and "sticks"... saturation)
First I thought it was the design..nope
It was the 1381/3503's ,these devices EXCEL in normal
operation.. but when they saturate, they really stick hard.
So , I resimmed mje 340/350's.. bang, fixed the problem.
also, if you put a .1uf cap across the Re of the CM so even
spikes from near saturation are gone..
In your amp (abomination) I just had to adjust Re's on
the VAS cascode to get symmetric clipping ,added that .1u cap
(your V- CM is the same as mine) all your cascodes helped
prevent saturation for the most part.
OS
lineup said:
AD797 second stage idea.
The idea is to use a floating mirror - gives at all output voltages same Vce across Q8 and Q9.
Lineup
Well, the idea behind floating CM is to present the same higher impedance on both sides of the LTP outputs, hence a very well balanced first stage at all working frequencies, hence a very high OLG, but unfortunately, only at LF.
The Achilles heel of this topology is VAS output impedance, which is high. Any attempt to load the VAS will ruin OLG.
It actually works really well only coupled with a very well designed output stage.
ostripper said:
I had similar problem with these same and some others models - from fairchild site
especially ...
I use LTSpice.
I was done small modifications in models:
*+ QCO =0.05
*+ RCO =50.1187
- comment these parameters or set to zero.
And small quote from KSC3503.lib:
* If QCO and RCO are set to zero, this model applies to
* the Gummel-Poon model LEVEL 1.
LEVEL=2 parameter is unrecognized by LTSpice anyway, so it seems You can sfely
comment it out, then all works fine.
lineup said:
Hi lineup,
I try to avoid complications if the results are not going to realize a substantial improvement. One area in this topology that I'm trying to avoid changing is the VAS. I don't want to upset the stability of this design - it is the biggest attraction.
roender said:
Thanks for the explanation roender,
What do you think of my latest schematic? I think I have the bugs worked out and I've taken it to it's limit in simulation. I need to build it to try out the changes that I've made and compare to the performance of the Abomination.
Can you see any glaring mistakes?
MJL21193 said:
MJL21193
Maybe you did not notice, but I said:
... maybe some time if when you wish
Meaning, put it into your knowledge bank and try it out some other time.
Now, I know you have got a new idea from me.
Would I expect a thank you .. maybe.
But I can not require any thanks from anybody, I know
There are so, so many people without any sense of gratitude on our forum.
Lineup
ostripper said:
Sorry OS, but that is not the case. I swapped the 2SA1381/SC3503 for the MJE340/350 in my version of the Frugal II and the clipping is actually worse.
Keep in mind that I reduced VAS current to ~2.2mA per leg to help the clipping instability problem before.
I have solved the gross clipping performance of this latest amp by going back to the LED as opposed the the rather ridiculous way I had it configured when roender commented on it. I expect the simulated clipping performance to match the real work amp - as seen earlier a shot of the Abomination clipping 5Vpp over input sensitivity.
wojtek5001 said:
Hi wojtek5001,
I use Multisim and I don't have any problem with the 1381/3503 models from Fairchild. They seem to be very good.
MJL21193 said:
I was talking just about problem that (as I think) Ostripper experienced, as hi said "sticky clipping" or something like that. I don't tell that these models are bad - they're
probably good, but just behaves strangely when used in VAS and amp is driven to clipping.
And that was tested only with LTSpice - on other software I have no idea.
To Ostripper: can You try my mods and tell if that helps?
Edit:
hi sad: "but when they saturate, they really stick hard"
O.K. I was look at Your picture with that "sticky clipping",
and must admit, that it looks very similar to my sims.
But in sims that happens regardless of frequency.
Even at 100Hz something "stick" for about 2ms
(I'll try to post some pictures tomorrow).
That sticking dissappears when i correct model as in my first post.
and must admit, that it looks very similar to my sims.
But in sims that happens regardless of frequency.
Even at 100Hz something "stick" for about 2ms
(I'll try to post some pictures tomorrow).
That sticking dissappears when i correct model as in my first post.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.