PDA

View Full Version : class A amp project for Mark



Ian Mackenzie
06-17-2007, 06:29 AM
I am not sure if this is the right forums (not Lansing diy) but here is a link to a project I am helping build with Mark (pasadena)

http://www.diyaudio.com/forums/showthread.php?s=&threadid=100807&perpage=10&pagenumber=1

Most of the technical stuff will no doubt be boring for you guys so I will post a few pics as well

The amp is rated at 100 watts rms nominal into 8 ohms per mono block.

This design is a diy version of the AX Passlabs amp series and uses the Pass Labs Super Symmetry and Aleph current source patents. This is a diy exercise so there is no infringment of the patent.

It is intended that this amp will ultimately be used to drive a set of XPL200's in biamp mode.

To start with there are some concept images.

Ian Mackenzie
06-17-2007, 06:35 AM
I decided on a square box for weight reduction and structural integrity. The square design also avoids the large rectangular box typical of this type of amp.

It measures about 260mm cubed or there abouts. The heatsinks are 100mm x 150 mm x 44 mm and are rated at 0.61 c / watt each. In use the amp will idle at about 45c in at an ambient temperature of 25c and draw about 250 watts off the mains

The transformer was specially made and is rated at over 625Va and has two 21 volt secondaries rate at 11.9 amps

This is over 264,000 uf (12 x 22,000 50 volt Panasonic) of filter capacitors used in a CRCRC smoothed DC supply.

There will be 16 IRF240 power fets per channel, in use each fet is biased to about 0.7 of an amp for a total of about 5.6 amps of bias.

I must say what seemed like a neat idea has turned out rather complex and a long job in the making. Numerous tools and jig and been either purchased or made for this type of fabrication.

The material is aluminium sheet (4mm angle), 3mm & 4mm plate and bar stock size of 12 x 20mm and lots and lots of drilled and tapped M4 and M3 holes.

Ian Mackenzie
06-17-2007, 06:42 AM
Here is the control pcb and associated mounting hardware.

Also some images of the pcb etching process, the powersupply simulation... and construction in progress.

Questions from interested posters welcome.

Ian

Uncle Paul
06-17-2007, 09:51 AM
Ian,
This is fascinating stuff for me. As always, you planning and execution is first class.

I'm planning a bi or tri amped 4344 inspired DIY using the MOX crossover and Aleph amps. Not sure if the LF amp will be Aleph, or if I want to use a different class there.

I'd go 4345, but have to unfortunately compromize due to size. Cabinet drawings are finished, parts are sourced, and construction has begun.

Electronics are phase 2 of the project. I have enough amps and crossovers to keep them playing while I complete the DIY electronics.

I've been pouring though the threads (including yours) in the Pass Labs forum at DIYaudio trying to make sense out of all the various versions and variants to determine which is/are the best for my application. It seems that the Aleph X, though 5 years old, is still pretty tough to beat. I'd appreciate any input you have on this.

I plan on integrating the crossovers into the amp pcbs to reduce the number of cross connects and related woes.

I learn something new everytime I read one of your posts, here or there.

Thanks,
Paul

Ian Mackenzie
06-17-2007, 01:49 PM
Hi Paul,

Enjoy your loudspeaker project.

The Mox sounds like a good plan if you are going the Aleph route.

It always brings out the most natural sound.

I will be incorporating a Mox in a Pass diy preamp as a followup to this project.

The std Aleph is a nice choice and is easier to build.

They have a more romantic presentation and work well with JBL's.

The X Aleph brings excitement and revelation from the energy and spatial detail in the presentation

Because of size and heat issues consider one of the better class A/B amps for the woofer.

If you are going diy have a look at the SKA GB 300 for the woofers or one the Anthony Holton aussiamps. Some of the class D amps are also interesting.

Ian

Guido
06-18-2007, 03:26 AM
The std Aleph is a nice choice and is easier to build.
They have a more romantic presentation and work well with JBL's.
Ian

Yes, Yes and Yes!!!

Ian Mackenzie
06-18-2007, 05:43 AM
I find it ironic there are no Yanks to speak of (except one) around here that use them.:barf:

The seem to prefer road re surfacing machines for amps by comparison...the yanks do love their roads I guess!;)
http://www.freepatentsonline.com/5080524.html


http://www.freefoto.com/preview/21-87-51?ffid=21-87-51&k=American+road+%2F+USA+road

Hoerninger
06-18-2007, 06:23 AM
Ian,

please go on and choose your own road, I would be pleased to follow.
Class As are real fine (and pretty warm ;) ).
___________
Peter

JBLRaiser
06-23-2007, 05:10 AM
It was a joke and a hint but speaks volumes for the demographics and VALS of this forum versus the enormous following over on diyaudio.com forums. I few soldering tips and your own your way.....

From a practical point of view (see 30 watt Aleph thead that won't break the bank so) anyone here can enjoy Hi end amplification versus the junk we so often see on Ebay. Franky I would never own an aged mains appliance, its just too dangerous.

I use a Dynaco ST400 for the low end, coupled with a Dynaco ST35 for the horns to my model 19's. Almost 40 year old gear. Any thoughts or comparisons with what you are doing. I'm considering something more technologically advanced.

Ian Mackenzie
06-23-2007, 05:24 AM
JBL Raiser,

I have never used the ST35 model so I can't comment.

But if the comments in these links are anything to go by the Model 19's horn compression drivers maybe the limiting factor. Might be worth looking at the newer versions on Great Plains Audio (Iconic)

At any rate the Aleph 60 or Aleph 4/2 would be a nice upgrade from the ST400. The bass response from these amps is quite amazing. You might even get away with running them full range and get a better result than attempting to bi amp (unless you are using an simple RC passive filter)

See Guido's Aleph 2 project.

Thanks for your post

http://www.tnt-audio.com/ampli/dynaco-st35_e.html

http://home.indy.net/~gregdunn/dynaco/components/ST35/index.html

Ian Mackenzie
06-23-2007, 05:26 AM
Today I did some more construction of the power supply smoothing filter banks.

This is quite a fiddly aspect of the project and I am pioneering as I go so its slow but steady progress.

Things have to be made to fit a certain way and there are numerous considerations. Temperature was only 5 degrees this morning before the heating at Area 51 kicked in. The aluminium almost to cold to touch.

Here we have the first two stages of the CRCRC filter nearly complete.

The 22,0000 uf 50 volt Panasonic capacitors are soldered and hot melt glued to a large pcb that will be supported by nylon stand offs on 20mm x 6mm bar stock with M3 hardware drilled and tapped throughout.

This construction forms a support frame that bolts to the heatinks on each side and provides further structural rigidity.

On the outside of the filter bank are the mini fin heatsinks that are attached by L section to the bar stock.

They draw heat from the power resisters. About 5 watts per heatsink. The power resisters will be thermally bonded and bolted in place with M3 hardware & tapped holes in the heatsinks.

The third stage of the CRCRC network will be located adjacent the main circuit pcb as shown and will be arranged using similar construction as above. This will be complete next week

A close up of the main pcb.

Deceptively simple but very elegant in execution.

Some of you may think this is overkill but as a result of the class A operation the power supply must sustain the voltage and current continously and be very reliable over the long term.

This power supply also has very low noise from the mains and ripple from rectification to ensure the best performance from class A operation.

I anticipate initial testing in about 2 - 3weeks

Uncle Paul
06-23-2007, 10:19 AM
Thanks, Ian. Looking forward to your update!

Ian Mackenzie
06-28-2007, 03:50 PM
Some more progress yesterday.

Here the intended view of the PS and main pcb.

I intend to test and match the power fets tomorrow and have it ready for testing within a fortnight.

Ian

pasadena
06-28-2007, 07:37 PM
Hey Ian,

Wow......That looks undoubtedly fantastic work. The innards look complex and meticulous.

I can see when you mean this beast is going to end up weighing a ton.

Shipping's going to be a challenge I must say.

Waiting for that blue light to flick into action. ;)

Cheers
Mark.

Ian Mackenzie
06-29-2007, 02:33 AM
http://www.diyaudio.com/forums/showthread.php?postid=1244445#post1244445

Ian Mackenzie
07-01-2007, 05:07 AM
More chunks of aluminium...

The diodes would not seat property on the little pcb heatsinks so I had to look for another solution.

Given the surface area of the lower outer panels I used a 20x 25mm solid bar between the diodes which will be thermally bonded to the 80 x 260mm panel.

This was non routine as the design had to be adapted to suit the circumstances. I am glad it turned out this way because the heatinking is far more substantial and the recifier will be located close to the filter bank which will minimise noise emmissions.

The bar was tapped with M3 holes for the rectifier pcb and M4 to attach to the panel.

Also an image of the soft start circuit and the Mitre saw.

A soft start circuit is used to moderate the inital high surge current from the large toriodal transformer and intial charge of the filter capacitors.

It consists of the time delay for a relay that closes after about 3 seconds. The current initally passes via a thermister that has cold resistance of about 10R (ohms).

Next week we will look at installing the switches and audio connectors.

Ian

Ian Mackenzie
07-07-2007, 11:35 PM
Worked on completing the router table today

Spent the day milling out some slots and drill some big holes with the big drill.

Guido
07-13-2007, 03:48 AM
Ian,
do you use these Kapton isolators?

Ian Mackenzie
07-13-2007, 04:22 AM
Up until now I have only used mica .

Ian Mackenzie
07-15-2007, 06:28 AM
The construction of the router table is nearing completion.

Sorry I forgot the camera.

I worked on the guide rail for the router today. The guide is a 1 metre aluminium extrusion with a 3/4 inch steel bar that is attached to the router base and runs the full length of the guide. I made a temporary base out of 6mm masonite for the initial trials.

The table has an overall length of 1 metre and an effective travel of about 500mm.

The idea is that the router will slide atop a substantial alumnium channel that forms a clamp to press down on the work (plate aluminium)

At each end of the channel there are large spring loaded hex bolts that will draw the channel down on the work. A length of 3mm hard neoprene has been attached under the channel which will squeeze against the work under pressure of the bolts.

There are also four guides to hold the channel in an extact lateral position E- W , N - S. Another set of guides in the form of a 500mm stainless steel ruler will be the work stops. The image below was taken in the early stages but you can see the channel construction

I saw a table by tool company today about 1/2 the size of this diy table and it was about $1100. The BOM so far is about $250 (less router)

Ian Mackenzie
07-21-2007, 06:27 AM
Not much to report here due to a bad dose of the flu.

Ian Mackenzie
07-28-2007, 09:38 PM
I am nearly ready to throw the big switch.

Completing the final fit out of switches and conenctors today.
Most of the infra structure is done.

Need to buy a cutter for the 24mm connector holes. A 5 minute job with the right tool.

Musing some ideas with Mark on the cosmetics.

Ian

Ian Mackenzie
08-02-2007, 10:04 PM
We looked some options including anodising but that was difficult to arrange in small quantities and with cut and machined work.

So after a lot of consideration I have arranged to get the exterior plates brushed and nickle plated.

Some of the work was dropped off at the plating works today so I should have something to show you next week.

I think it will look look quite nice. :)

Ian

Ian Mackenzie
08-05-2007, 03:22 AM
I made more progress on the amps this weekend.

The usb uplink cable of the Nikon broke so I will attempt to descibe what has happening behind he scenes.

1. Yesterday I cut out the holes on the rear 26 x 8cm aluminuim panel for the Speakon and Canon 3 pin sockets and drilled the M3 mounting holes for the connecters. The panel will go to the plating works sometime this week.

2. The amps were striped down today to nut and bolt level and the remaining extrusions are now ready to be sent for plating this week.

3. The heatsinks were polished and cleaned ready for final mounting of the power fets and associated Pcbs. This is to ensure a smooth flat surface free of oxide build up. Everything will be carefully cleaned and degreased again before final assembly.

4. I also made a nifty Jig for bending the power fet legs to precise measurements so they pop through the pcb holes in just the right position.

5. Final wiring was soldered to the power Fet pcbs earlier in the week and are ready for installation.

6. The final power filter for the CRCRCpcb was fabricated and I have managed to fit it neatly under the main pcb. The benefit of this is the whole assembly represents a bolt in module and the filter capacitors are within an inch or so of the main pcb and the output stages. There is 44,000 uf on the final filter bypassed by 2,200 uf on reach supply rail. The supply is rated at 12 amps continous current +-25 volts.

7. I managed to source some TO3P Mica washers during the week from a repair centre. I am certain they wondered why I counted out 32 washers..they are now out of stock. Mica is now quite hard to obtain. They are by far the best thermal washers.

8. In order to expedite the final assembly the cutting of the top and bottom panels will be out sourced to a shop. Completion of the router table will be pursued at a later stage when I have more time.

So its all coming together at last.

In the attached image taken at an earlier stage looking down into the works you can see the primary capacitor bank.(44,000 uf, 0.235R 30 watts, 44,000uf, 0.235R 30 watts). The final stage of capacitors mentioned in item (6) will be mounted inverted under the bar stock bracket holding the main pcb at the top of the enclosure. So there will be a central core if you like that will be the power supply. Since that pic was taken the brackets have been adjusted and the top bracket has been raised about 15mm and the side vertical edges trimmed in the Makita Mitre saw so they are flush with the side vents which extend the full perimeter of the chassis.

Fitting everthing in has been fun but a painstaking and tedious task at times and as a result quite time consuming.

While the capacitors are rated at 105c (for a given ripple current) there should be sufficent temperature/pressure differental between the bottom and top vent channels to ensure adeuqate convection cooling. The amp is designed to run at 46-50C (in an ambient temperature of 25-30c) without forced air cooling.

This is the idle temperature and the optimum operating point for best measured performance. The amps will take about an hour to reach full temperature. The gain of the fets and some their characteristics change with increased temperature so it is important that the bias and other tweaks be done at full operating temperature.

I would expect the subjective performance to be at its peak in well under an hour. Forced air cooling a frowned upon in class A amplifers because the ambient background noise masks important low level resolution. Even the quietest fans are audible so we cool these amps with sheer heatink real estate. There is collectively 800mm of 150mm high heatsink in each amp.

In the early stages I talked to Mark about his situation and as a result designed the amps for the Perth WA summer. So on a really hot day (inside the house) with an ambient temperature of 30c the amps would be running at 51c. That is hot to touch but not hot enough to burn someone (Mark's kids). 55c is on the threshold of pain and 58c is too hot to keep your hand on the heatsinks.

Typically temperatures in Perth can soar to the high 40's in the summer and allowing for air conditioning I would not expect anyone to be in a listening room at 30c ambient temperature running a pair of class A amps pumping 500 watts into the room but we design this into the amps. A thermal breaker opens at 75C.

The biasing and absolute Dc offset can take some time to settle with this design, typically an hour. Absolute dc offset is the potential difference between the positive and negative loudspeaker outputs and ground. Because the positive and negative outputs float above ground (the amp outputs are balanced SE) in normal operation we provide a load of between 50-100 ohms to help control this offset at turn on. Some common mode feedback is alse used to help control this offset.

The loudspeaker does not see this offset and typically its 5 volts at initial turn on and then slowly settles to within 100mv after an hour. The actual DC offset is only small, typically 10-20 mv at all times.

So you can see the operation of the amplifier is closely allied to the physical parameters and we measure key indicators when certain conditions are me.

The final adjustments will be done over several days.

Ian

Krunchy
08-05-2007, 12:52 PM
Wow, I stand humbled.:spchless:

Audiobeer
08-05-2007, 06:33 PM
Great read Ian. can't wait to hear more!

Ian Mackenzie
08-05-2007, 07:04 PM
Thanks,

I can't wait to here it go.

Need to buy a new usb cable for the camera!

Ian

merlin
08-06-2007, 03:07 PM
One thing's confusing me sat here in England Ian.

Where's the tubes?

Ian Mackenzie
08-08-2007, 06:03 AM
Here are some recent images of assembly and fabrication.

My favourite tool..the big drill. Image of 24mm hole's being drilled for the connectors. Out of frame is a large clamp holding the work.

Next image of the rear panel (un plated and finished at this stage.

The pcb assembly modules showing the final filter capacitor storage stage under the main pcb.

.

Ian

Ian Mackenzie
08-08-2007, 06:06 AM
Here we are again,

Another angle of the pcb modules and the heatsinks stacked up following polishing and cleaning.

Ian

richluvsound
08-08-2007, 06:15 AM
Hi Ian ,

nice work . I heard Guido's DIY pass at the weekend. If those are half as good then Mark is a very lucky man.

Rich

Hoerninger
08-08-2007, 06:29 AM
My favourite tool..the big drill. Image of 24mm hole's being drilled for the connectors.

Great pictures - great posting, Ian.
How do you manage the rectangular holes?
___________
Peter

Ian Mackenzie
08-08-2007, 02:09 PM
Great pictures - great posting, Ian.
How do you manage the rectangular holes?
___________
Peter

With great difficulty.

I really would like a milling machine. I basically mark out the hole and vary carefully drill it out. I then use a fine saw (tiny) and cut through all thedrills holes . Then carefully file the edges using a lubricant. Each hole takes an hour.

Ian

Hoerninger
08-08-2007, 02:22 PM
Ian, that is the way I have learned it.
I feel with you. :D
___________
Peter

Ian Mackenzie
08-08-2007, 02:24 PM
Hi Ian ,

nice work . I heard Guido's DIY pass at the weekend. If those are half as good then Mark is a very lucky man.

Rich

Thanks Rich,

It will be interesting. This design, the X Aleph is a balanced (Super Symmetry tm SE) version of the Aleph . Basically that means lower distortion (2nd harmonic) and more dynamics and better bass (particularly lower power versions).

Its like an Aleph on steriods.

Here is a link to my first X Aleph I built some years ago

http://www.passdiy.com/gallery/alephx-p1.htm

The Aleph 2 is still my favourite Aleph and its easy to make.

I also have a Zen V9 and a First Watt F4 I will be building shortly.

Ian

Ian Mackenzie
08-10-2007, 08:30 AM
Just finished matching all the power Fet..took 3 hours. They are all within the span of 4.18 - 4.28 VGS.

Had a breather at about midnight watching Bruce Willis and some Dawwggy diving competitions in a pool in the city on a N.Y tonight show. That as different.

Ian

Robh3606
08-10-2007, 10:07 AM
Hello Ian

It's Dawwwgg

Nice Amp!!

Rob:)

Ian Mackenzie
08-11-2007, 02:01 PM
The plating looks very good. Basically everything with the exception of the heatinks have ben plated. Difficult to photograph in harsh lighting

Almost fully re-assembled the chassis. It took about 4 hours to reassemble both chassis from my engineering frame drawing..each panel is coded.

Looks the part.


Ian

Ian Mackenzie
01-01-2008, 02:40 AM
Hello and Happy New Year.

Been busy in the kitchen (in joke in the Pass Diy divison)

Attached are some images of the project in the final stages.

For the most parts its gone according to plan although it runs a little warmer than I would like. The wiring shown is for testing only.

Some data.

The transformer is 626 va, 21 volt secondaries @11.9 amps, supply capacitors 22,000 @ 50 volts x 12, high speed Hexfet rectifiers, 16 IRF240 power fets, total dc bias 5.33 amps, maximum power 120 watts into 8 ohms

richluvsound
01-01-2008, 04:08 AM
Happy Year Mate !!!!

Ian,

they look brilliant ! Are you happy with the overall sound ? Do they demonstrate the Pass qualities we worship ?

Rich

Ian Mackenzie
01-01-2008, 04:35 AM
The design is an upscaled version of this:

http://www.passdiy.com/gallery/alephx-p1.htm

See the last paragraph..I wrote that several years ago.

Ian

Guido
01-01-2008, 04:55 AM
VERY NICE Ian.

Did you manage to get rid of the transformer hum?

Ahhh Panasonic FC's

I used Elna Silmics this time.

Ian Mackenzie
01-01-2008, 12:10 PM
Hi Guido,

It was something else in the room oddly enough.

Now almost completely silent.

The key is to use a transformer 7 x the power output and have it wound for now noise. My transfomer as quad secondaries

I have some Elna Silmics, I will try them.

Its like small stove when it switched on!

Anyone for a Barbi...

Ian

Ian Mackenzie
01-13-2008, 02:05 AM
I bit more progress.

After testing for two weeks I connected them to the JBL.

The sound, in a word "spooky".

There are some other things I need to sort out and will post details next weekend.

Ian

Guido
01-13-2008, 04:02 AM
The sound, in a word "spooky". Ian

Spooky good, or spooky bad :p

Ian Mackenzie
01-13-2008, 01:45 PM
Limitless...like Forbidden Planet.

Robh3606
01-13-2008, 02:48 PM
Make sure the ID doesn't get the best of you. You don't want to terrorize people while you sleep! Looks great.

Rob:)

Ian Mackenzie
01-20-2008, 01:34 AM
Rob,

ID has come through the wall literally.

I had an opportunity to wind it up this afternoon while the rest of the neighbourhood was recovering from last night’s tennis.

If the ID were the sound it can only be described as watching the Queen Mary being spun into a blur on a dime and stopped instantly!

The dynamic contrasts are spectacular and it shows in just how accurate the JBL's portray this aspect The degree of realism in this respect is quite uncanny.

These qualities extend to a sense of tone, of depth and dimensionality that cannot be expressed in words.

I imagine a 1500AL user would be beaming at the thought.

Put simply its a music lovers amp.

Be it heavy metal, rock , jazz or acoustic solo.

There is scope for some tweaking and following consultation with the The Master and some other diyaudio people we plan to try a parallel cascoded P channel Jfet front end swap.

Ian