♫♪ My little cheap Circlophone© ♫♪ - Page 25 - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 5th September 2011, 04:10 AM   #241
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
Direct comparison was becoming more difficult as the design
drifted away in a few details that weren't necessary to differ.

I put drivers back in, as they were needed to handle 4 ohm.
Sorry not the same BD140's. I kinda painted myself into a
corner where only an NPN would work, 2n3019 volunteered.

I also lowered the quiescent, so it crosses more similar to
Elvee's Circlophone. Could use a little more tweaking here,
as I think I still have my reserve current set slightly higher.
Attached Images
File Type: gif converge.gif (30.7 KB, 406 views)
Attached Files
File Type: txt converge.asc.txt (9.5 KB, 10 views)
  Reply With Quote
Old 5th September 2011, 11:32 AM   #242
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
Something ugly happening in this output stage now in hard clipping, ???
Quick and dirty fix is to soft clip the VAS, rather than the global loop.
Just disconnect lower end R15, and reconnect to the wire on it's right.
Soft clip is no longer symmetrical, but the hard problem was worse...
Just till I figure some better way to fix it... I may have output drives
biased a little lower than I should.
  Reply With Quote
Old 5th September 2011, 12:08 PM   #243
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
Reconnect R15 to GNF, the way it was.
Add BAT54 from the top of R15 to VAS.
Cathode end toward VAS.

Still not perfect, that extra diode drop
in the lower Darlington is where it really
goes to hell in a handbasket. I may yet
relocate VAS clamping to that area...
  Reply With Quote
Old 5th September 2011, 12:17 PM   #244
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
This one looks quite nice and well behaved.

There seems to remain some work to do in the dynamic behavior/large signal department, as the waveform at 20KHz looks visibly distorted, but at this stage, actual physical tests would be needed to ascertain the effect, and refine the design if necessary since the sim may not be quite reliable for this aspect.

Compared to the Circlophone, it is somewhat simpler (3 small signal transistors less) and has slightly lower performances, but the most important differences are elswhere:
-It is built around Ken's pet obsession with schottky's to shape the crossing.
-It has a less deterministic control of the quiescent current and output crossing.
The Circlophone servoes very tightly those parameters in the low power/control circuits, whereas "Converge" works directly at the output level, utilizing the I/V characteristic of the schottky's as a template.

Both approaches certainly have their merits.
In the Circlophone I wanted to completely eliminate the dependency on components, thermal couplings and ambient temperature: you can use any schottky's, any power transistors, any low power transistors, it always work identically, any spread being absorbed by the servo. In fact, you could even include one germanium in the mix, it would still work.

In "Converge", the quiescent current will depend on the type of schottky, their temperature, the saturation current of the sense transistor, etc.
But it will give a special "taste" or colour" to the output stage's transfer.

It would be a good idea to launch a dedicated "Converge" builder's thread to put these ideas to the test.


Back to the Circlophone.....

I have built one more variant, this time with BD249's as ouput devices.

The purpose of this version was to test the design under extreme conditions, in this case a very low impedance load.
Some "exotic" speakers have an impedance <1Ω, and few amplifiers are capable of driving such a load directly, without a transformer.

I made a test on a 0.5Ω load.

This pushed the transistors to their limits, at >25A Ic, but the Circlophone remained steady as a rock.

The sim uses TIP35's, as I couldn't find models for the BD249C, but they are very similar.

This illustrates how tolerant, flexible and healthy the design is.

Of course, the THD is increased compared to 4Ω, but it remains quite acceptable, especially taking into account the fact the transistors are really pushed to their limits (and beyond).
Attached Images
File Type: gif CirclophoneLZ.GIF (60.1 KB, 389 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 5th September 2011, 03:04 PM   #245
Bigun is online now Bigun  Canada
diyAudio Member
 
Bigun's Avatar
 
Join Date: Jan 2009
Blog Entries: 2
Well, I have a dirty great heatsink with 2N3055's all over it - the plan is a stereo JLH 10W class A. But there could be room in the chasis for another pcb to try out one of these circuits. I don't think it will be soon, but I'm starting to see the potential...
__________________
"The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.
  Reply With Quote
Old 5th September 2011, 05:28 PM   #246
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
Greasy, grey, dirt cheap components are a great challenge for design to overcome.
Make it work as well with junk as it would with unobtanium on three week lead time
and a thousand piece minimum order. It can be done...
  Reply With Quote
Old 5th September 2011, 05:48 PM   #247
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Here are some pics of the above-mentioned version:

The feat is made possible thanks, amongst other things, to the absence of emitter resistors: in a standard topology, with a 0.5Ω load impedance and 0.22Ω emitter resistors, practically half the equivalent of the useful amount of power dissipated in the load is wasted in the resistors.

In the Circlophone, the only obstacles to the current flow are high-current schottky's, and the transistors themselves.
The BD249 are rather cheap and crappy, and with better devices, loads even lower could be handled with a better efficiency.

But the Circlophone is not simply about brute force.

In a completely different domain, at a gentle power level and on more "normal" loads, the performance is on the par with a good, pure class A amp:
Attached Images
File Type: gif LopoCirclo.gif (41.6 KB, 355 views)
File Type: jpg 104_5024.JPG (303.4 KB, 360 views)
File Type: jpg 104_5022.JPG (482.1 KB, 365 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 5th September 2011, 06:11 PM   #248
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
Quote:
Originally Posted by Elvee View Post
There seems to remain some work to do in the large signal department.
Yeah, yours too... (1V Input)
Do I see transistors turning off?
Q1 Q2 Q3 Q4 Q5 Q6 Q9 Q10 Q12

Laugh again my non-determinstic clip and crossings?!?
Hold that egg in too tight a grip, this what happens.
Attached Images
File Type: gif 1VIN.gif (19.0 KB, 193 views)
  Reply With Quote
Old 5th September 2011, 07:53 PM   #249
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
I can do deterministic too.

Abuse a threshold narrow enough, Schottkys must cross class C.
Then set your deterministic class A crossing with two resistors.
Plus always-on bypass current...

Near as I can tell, its bias logic exactly the same as circlophone.
Though can't quite set my bias as low, on account of bootstrap.

Input here was 0.7 so I could compare apples to Circlopone M's..
Attached Images
File Type: gif deterministic.gif (45.2 KB, 112 views)
  Reply With Quote
Old 5th September 2011, 09:35 PM   #250
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by kenpeter View Post
I can do deterministic too.

Abuse a threshold narrow enough, Schottkys must cross class C.
Then set your deterministic class A crossing with two resistors.
Plus always-on bypass current...

Near as I can tell, its bias logic exactly the same as circlophone.
Though can't quite set my bias as low, on account of bootstrap.

Input here was 0.7 so I could compare apples to Circlopone M's..
Sorry if I offended you, that was certainly not my intention.

Since you chose this thread to display your ideas, I gave my opinion, and made some comparisons.

I made comparisons, but I don't think you interpreted them correctly.

I just meant that you solve the problem using an "open loop" approach, where my solution is a more "closed loop" one, where things are more tightly under control.
This is not a value judgement, more the observation of a difference.
Both ways have advantages of their own, and it is impossible to decide in absolute terms which one is the best, if there is such a thing.

To be honest, I find the final version of the Circlophone looks quite clumsy, inelegant and over-complex.

I started with something cleaner, more symetric and more elegant.
But I reworked it, made many actual tests, and arrived at the painful conclusion that a rather messy circuit was ultimately the optimum solution.

In the end, I settled for something rather ugly, but effective.

I wanted to tightly control the quiescent current and crossing regions, as I said, because that's where most of the problems associated with conventional class AB occur.

I never said I wanted to control the clipping conditions in the same way: that's one of your own preoccupation, which I don't necessarily share.
I just wanted to make sure the "outside" behavior was decent, no more.
I haven't included "clean" or "soft" clippers: I just made sure the clipping behavior, as seen from the load remains clean, but without attempting to soften or improve it.

As you have pointed out, the slow transistors struggle to get out of saturation, and they are "nudged" by their counterpart to do so, but the process remains completely transparent to the load: the clipping is clean and the recovery is clean too.
See the pic of the output voltage during clipping (the actual waveform is identical to the sim).

There are spikes of current in the OP stage, but they remain moderate, and perfectly manageable by the OP transistors, and this has been abundantly tested in reality.
I didn't want to control further the clipping behavior, that's my design choice, you made another one, you are free to do so, OK why not, but I'm also free to do mine: I consider clipping should be exceptional, and that a clean behavior is sufficient, period.


If I said the dynamic behavior of your circuit could be improved, I didn't mean the clipping: I meant slew rate and power bandwidth.

As I said, clipping is a secondary consideration for me, but I think a decent behavior at 20KHz does have some importance.

I am not an extremist in that matter, unlike many on this forum, but I think a decent margin is desirable, and achievable without too many difficulties, even using rather crappy components, as I do.

If you examine the power bandwidth of your amplifier, you will see that there is room for improvement.

I was cautious in what I said, because the physical version of the Circlophone did much better in this respect than the sim, and that was verified with all 2N3055 brands, without exception.

But where there are obvious shortcomings, they should be remedied.

One more thing:
I do not want to look patronizing, or hurt your feelings
If you think/feel I am, this is in no way intentional.

I think you are one of the most creative members of this forum, and I think your latest creation, the active tail control for the LTP is great.
I have even tried to find ways to include it in the Circlophone, but unfortunately, the two principles look mutually exclusive.

But it will stay in a corner of my mind, available for a future use.
Attached Images
File Type: gif Circloclip.gif (38.4 KB, 132 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Need help with OB design - cheap, cheap, cheap djn Multi-Way 38 30th May 2010 06:26 AM
Scosche SPL meter: Cheap find with potential, or cheap junk? theAnonymous1 Everything Else 5 11th October 2006 03:40 AM
Alpine cda-9831 vs. 9833 vs. 9835 cbecker33 Car Audio 3 10th December 2004 12:47 PM


New To Site? Need Help?

All times are GMT. The time now is 11:46 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2