Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Solid State Talk all about solid state amplification.

SLAPS for SLAM!
SLAPS for SLAM!
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 13th October 2019, 04:02 PM   #1
solhaga is offline solhaga  Sweden
diyAudio Member
 
Join Date: Jun 2009
Location: Värsta Götaland
Default SLAPS for SLAM!

The SLAM! AMTs I built have a high efficiency (100 dB/ 1 W / 1 meter), so I don't need that many Watts to drive them.

I'll be quite satisfied with 4 Watts, the AMT's resistance is 6 ohms, so 4.95 Vrms is sufficient.

The AMTs also have an outstanding resolution, but they are at the same time merciless in revealing any amplifier's misconduct.

My current 25 W class A amplifier is no exception.

I've been suggested to give up on complex amplifiers and instead try something simpler and with no feedback at all.

This is the SLAPS (Split load phasesplitter amplifier) amplifier I've been suggested by Circlomanen:

Click the image to open in full size.

It amplifies only 2 times, so I need a pre-amplifier with 12 times amplification.

More to come...

Last edited by solhaga; 13th November 2019 at 06:57 PM.
  Reply With Quote
Old 13th October 2019, 05:02 PM   #2
solhaga is offline solhaga  Sweden
diyAudio Member
 
Join Date: Jun 2009
Location: Värsta Götaland
So letting LTSpice simulate with a bias set to give an Id of approximately 3 amps gives:

Click the image to open in full size.

THD is -66.0 dB mostly second harmonics.
Third is at -87.5 dB.
  Reply With Quote
Old 13th October 2019, 05:16 PM   #3
solhaga is offline solhaga  Sweden
diyAudio Member
 
Join Date: Jun 2009
Location: Värsta Götaland
So what about reality?

Coils:

Click the image to open in full size.

Preamp:

Click the image to open in full size.

MOSFET:

Click the image to open in full size.

Power supply:

Click the image to open in full size.

I found that the MOSFET got really hot, so I got a bigger heatsink and also put a smaller one on the top of the MOSFET.

Click the image to open in full size.

I connected a 8 ohm dummy load and let REW measure it using my EMU0402:

Click the image to open in full size.

Not that nice as the LTSpice simulation, but the 2nd and 3rd relationship seems to be correct.
Now the wires is not ideal and I don't what the power supply contributes, but I'd say it is looking good so far.
  Reply With Quote
Old 13th October 2019, 05:34 PM   #4
solhaga is offline solhaga  Sweden
diyAudio Member
 
Join Date: Jun 2009
Location: Värsta Götaland
Time to measure with the AMT using REW and UMIK-1.
This is the middle range AMT, so it rolls off at 13 kHz.

SPL with a 1.1 dB matching:
Click the image to open in full size.
No difference to mention.

Distortion with the complex amplifier:

Click the image to open in full size.

Distortion IRFP054:

Click the image to open in full size.

Both, 2nd:

Click the image to open in full size.

Both 3rd::
Click the image to open in full size.
Here I am a little bit surprised that the two amplifiers behave almost the same.
But it could the AMT distortion that takes over.

Both THD up to 9th:
Click the image to open in full size.

2nd order is the dominant part as we have seen earlier.

I'd say that the simple amplifier looks promising considering the state of the circuitry.
I need four amplifiers as there's a treble AMT as well and they are all driven active.

But before I start to build proper circuitry and proper power supply and casing and so on, I think that there are room for improvements.
An Id of 3A makes the MOSFET really hot, I guess it will not last long.
The bigger heatsink is at 72 degrees Celsius and the smaller on top of the MOSFET is at 84 degrees Celsius.
For sure, the MOSFET's leads can be soldered to a bigger copper area, but it is still to hot.
And what about those 2nd distortion levels?
  Reply With Quote
Old 13th October 2019, 05:54 PM   #5
solhaga is offline solhaga  Sweden
diyAudio Member
 
Join Date: Jun 2009
Location: Värsta Götaland
Default Four is better than one

So I thought of have four MOSFETs in parallel:

Click the image to open in full size.
The resistor R8 and up is used to be able to measure current.
I changed the bias to give 1.5 A through each MOSFET; I think that should give a more normal temperature.
So the current through the coils is 6A.

LTSpice have no trouble with this:

Click the image to open in full size.
The distortion is down to THD -79 dB (3rd at -105,7 dB!) and I reckon that noise levels are down as well.

So now comes my question:

But how come the bias now is higher than with the single MOSFET one above?
It corresponds for sure with the Id of 6A for one MOSFET, but there are now four.
Is the LTSpice simulation correct?
  Reply With Quote
Old 14th October 2019, 03:44 AM   #6
sgrossklass is offline sgrossklass  Germany
diyAudio Member
 
Join Date: Nov 2004
Location: Germany
Remember what Einstein said about "as simple as possible, but no simpler"? Well, this circuit qualifies as "too simple" in my book. You've got a triangle of complexity - performance - power consumption going on, and in this case this means that by the time you get acceptable performance out of it, it'll be a real power hog. The only thing you've gained from this is an amplifier limiting the driver's distortion rather than vice versa. It is rather clever in that the coupled inductors cancel out magnetic flux from DC current (perfectly in theory, more or less imperfectly in practice).

LTspice inductor models do not model ferrite-related nonlinearity by default, and I'm not sure how far I'd trust MOSFET models for accurate modelling of distortion.

Also, there is no such thing as 4 exactly identical transistors in real life - current sharing resistors are probably a must.

If you insist on something simple with inductive load, here's what I would do:
Click the image to open in full size.
(supply rail bypassing not included)

This, however, would require a rather substantial gapped choke. If the inductor doesn't work out, a transistor CCS may be more attractive:
SLAPS for SLAM!-opamp-se-mosfet-buffer2-png
Or perhaps more elegant:
SLAPS for SLAM!-opamp-se-mosfet-buffer3-png

While you're obviously free to increase gain if needed, I've left it at 4.3 now, since I can't imagine why you would need a gain of 24 (28 dB) in the power amp for a 100 dB/W/m driver. I mean, you'd probably be coming out of the crossover at up to 1 Vrms (give or take), with a noise floor that may not be too much below -104 dBV (-110 dB re: 2 Vrms), depending on your DAC. You might need 16 dB, but but probably not any more.

I picked an IRF530 because good MOSFETs for linear applications tend not to be low rDSon types, quite the contrary.

The 3rd circuit still isn't ideal in that current draw alternates between the positive rail and the load, while negative rail current is constant... you'd have to make this a single-supply circuit with all-negative supply, but then you've just swapped signal currents going through a power filter capacitor for signal currents going through an output coupling capacitor... happier power supply, less happy load, not to mention the whole dealing with power-on/off transients business.
Attached Images
File Type: png opamp-se-mosfet-buffer2.png (15.4 KB, 1012 views)
File Type: png opamp-se-mosfet-buffer3.png (14.6 KB, 1002 views)
Attached Files
File Type: asc opamp-se-mosfet-buffer-3.asc (3.9 KB, 10 views)
  Reply With Quote
Old 14th October 2019, 08:41 AM   #7
Circlomanen is offline Circlomanen  Sweden
diyAudio Member
 
Join Date: Sep 2004
Location: Göteborg
Quote:
Here I am a little bit surprised that the two amplifiers behave almost the same.
But it could the AMT distortion that takes over.
Yes. The AMT own distortion will dominate over the amplifiers level of distortion.

You should make a distortion-measurement comparing the total level of second harmonic with the AMT coupled with the leads reversed to the amp.

"Triodity" to cancel loudspeaker distortion
See this thread in the Pass section of this forum.

I would not worry about that low level of second harmonic. I am more worried about transient and dynamic stability while driving a reactive load like a loudspeaker. A steady state or swept sine wave measurement does not say that much about the sound quality of an amp.
__________________
We live in a horizontal world, why use vertical topologies???
  Reply With Quote
Old 14th October 2019, 08:53 AM   #8
Circlomanen is offline Circlomanen  Sweden
diyAudio Member
 
Join Date: Sep 2004
Location: Göteborg
I prefer the much more stable and direct feed forward error correction in the amp you are building above using a global negative feedback for low distortion and low output impedance.

The feed forward error correction is stable and useful up to 1,5 MHz and does not need any filtering to avoid HF-instability while driving a reactive load through long cables.

The simplicity of the amp coupled with the use of feed forward error correction does give a very stable and nice behaving amp while driving highly reactive loads with non-linear signals.

Quote:
But how come the bias now is higher than with the single MOSFET one above?
4 parallel devices without any substantial current sharing resistors will have 4 times the transconductans of a single device.

If you want to use parallel devices you will need to match them closely and use current sharing resistors of a higher value then you have in your simulation. 0,22 - 0,5 ohm or so.

But I would prefer a lower voltage from the power supply to keep the single device cooler and I would not be worried about some small amounts of second harmonics.
__________________
We live in a horizontal world, why use vertical topologies???
  Reply With Quote
Old 14th October 2019, 03:09 PM   #9
solhaga is offline solhaga  Sweden
diyAudio Member
 
Join Date: Jun 2009
Location: Värsta Götaland
Thanks sgrossklass for your answer, but the simple circuit is what I'll go for.

Quote:
Originally Posted by sgrossklass View Post
Remember what Einstein said about "as simple as possible, but no simpler"? Well, this circuit qualifies as "too simple" in my book. You've got a triangle of complexity - performance - power consumption going on, and in this case this means that by the time you get acceptable performance out of it, it'll be a real power hog. The only thing you've gained from this is an amplifier limiting the driver's distortion rather than vice versa. It is rather clever in that the coupled inductors cancel out magnetic flux from DC current (perfectly in theory, more or less imperfectly in practice).
See Circlomanen's answer.


Quote:
Originally Posted by sgrossklass View Post
LTspice inductor models do not model ferrite-related nonlinearity by default, and I'm not sure how far I'd trust MOSFET models for accurate modelling of distortion.

Also, there is no such thing as 4 exactly identical transistors in real life - current sharing resistors are probably a must.
Good, I've added 0.47 ohm resistors.


Quote:
Originally Posted by sgrossklass View Post
While you're obviously free to increase gain if needed, I've left it at 4.3 now, since I can't imagine why you would need a gain of 24 (28 dB) in the power amp for a 100 dB/W/m driver. I mean, you'd probably be coming out of the crossover at up to 1 Vrms (give or take), with a noise floor that may not be too much below -104 dBV (-110 dB re: 2 Vrms), depending on your DAC. You might need 16 dB, but but probably not any more.
Well, +28 dB is what my current amplifiers have in gain.
It'll be an easy thing to adjust if I find noise levels too high.
I have no trouble with the noise levels with my current amplifiers.

Quote:
Originally Posted by sgrossklass View Post
I picked an IRF530 because good MOSFETs for linear applications tend not to be low rDSon types, quite the contrary.

The 3rd circuit still isn't ideal in that current draw alternates between the positive rail and the load, while negative rail current is constant... you'd have to make this a single-supply circuit with all-negative supply, but then you've just swapped signal currents going through a power filter capacitor for signal currents going through an output coupling capacitor... happier power supply, less happy load, not to mention the whole dealing with power-on/off transients business.
See Circlomanen's answer.
  Reply With Quote
Old 14th October 2019, 03:17 PM   #10
solhaga is offline solhaga  Sweden
diyAudio Member
 
Join Date: Jun 2009
Location: Värsta Götaland
Thanks Circlomanen for your clarifications.

Quote:
Originally Posted by Circlomanen View Post
Yes. The AMT own distortion will dominate over the amplifiers level of distortion.

You should make a distortion-measurement comparing the total level of second harmonic with the AMT coupled with the leads reversed to the amp.

"Triodity" to cancel loudspeaker distortion
See this thread in the Pass section of this forum.
I'll change the polarity between the amplifier and the AMT and compare the measurements.

Quote:
Originally Posted by Circlomanen View Post
I would not worry about that low level of second harmonic. I am more worried about transient and dynamic stability while driving a reactive load like a loudspeaker. A steady state or swept sine wave measurement does not say that much about the sound quality of an amp.
Right!

Quote:
Originally Posted by Circlomanen View Post
I prefer the much more stable and direct feed forward error correction in the amp you are building above using a global negative feedback for low distortion and low output impedance.

The feed forward error correction is stable and useful up to 1,5 MHz and does not need any filtering to avoid HF-instability while driving a reactive load through long cables.

The simplicity of the amp coupled with the use of feed forward error correction does give a very stable and nice behaving amp while driving highly reactive loads with non-linear signals.
Right again!

Quote:
Originally Posted by Circlomanen View Post
4 parallel devices without any substantial current sharing resistors will have 4 times the transconductans of a single device.

If you want to use parallel devices you will need to match them closely and use current sharing resistors of a higher value then you have in your simulation. 0,22 - 0,5 ohm or so.
0.47 ohm current sharing resistors it is.

Quote:
Originally Posted by Circlomanen View Post
But I would prefer a lower voltage from the power supply to keep the single device cooler and I would not be worried about some small amounts of second harmonics.
I'd like to keep the higher voltage as I like to have a little headroom.
Besides, the amplifier might come it use where more power is needed.
  Reply With Quote

Reply


SLAPS for SLAM!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

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
SLAM solhaga Planars & Exotics 54 12th April 2018 05:49 PM
SDX10 tips for slam? caper26 Subwoofers 8 14th June 2012 11:51 PM
slaps = acoustic lever? jbell Subwoofers 3 18th February 2009 10:17 PM
Email Slam. jleaman Everything Else 5 18th December 2003 12:49 AM
Wilson Audio Grand Slam X1 - Clones Percy Erasmus Multi-Way 13 30th April 2002 01:50 AM


New To Site? Need Help?

All times are GMT. The time now is 04:31 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Copyright ©1999-2020 diyAudio
Wiki