Aleph-X builder's thread.

I bought a case of the 680 VA toroids for 20 bucks a piece and the variac for 50 bucks...all surplus of course. There was a guy in Penn. that was selling the same toroid...I think he also bought them from the same place here in SLC that I did but his price was ALOT higher than 20 bucks. He advertiised them someplace on this site. None left at the surplus place but there are other surplus places that have usable toroids for this project. Check out #3851 for instance at this site..... Just about perfect!
http://www.elect-spec.com/surplusz.htm

Mark
 
Hi koolscooby,

From what I've read so far, I think your power supply may be a part of the problem. You should use a split supply with ground, rather than a single supply. However, you can do some initial tests with a single supply and voltage divider to provide the ground...

If we keep the operating currents low, a two-resistor voltage divider between the rails will give you a close enough approximation of ground. To help reduce ground current, use J1b instead of J1a (I recommend J1b anyway). Next, disconnect R7, R9, R36 and R38 to disable the output transistors. It wouldn't hurt to disconnect R46 and R47 too. Now we can concentrate on making sure your diff pair works properly.

Check the value of R23 and R25 with a multimeter, as they may have been cooked when you had C7-C10 jumpered.

Now, with everything connected as above, you should see 9V across D1, and roughly 5.5-6V across R23 and R25. With some signal input, you should get a nice signal between the drains of Q5 and Q7. If you've made it this far, then the diff pair is working just fine, and we can move on to bringing up the output stages.

By the way, you leave Q12b in... putting a jumper across Q12a simply disables it, leaving R17 to set the current through D1. If R17 is 1.5K, then this should work fine with J1b and with or without Q12 jumpered.
 
Chad, thanks very much for the steps, i'll try and do that stuff tomorrow night, or on the weekend, i was out on a conference this week, and am now back and getting unpacked and the crazy things that deal w/ that ;)

question: what resistor should i place between the -15 and ground and +15 and ground? i'm not aware on how to hook something like that up

Thanks everyone
 
For +/-15V rails, a pair of 1.5K 1/4W resistors should do just fine, but note that the ground split will be sensitive to ground currents. Double-check with a multimeter to see if you're getting reasonably equal voltages across each resistor. You can even go so far as to add a 47uF or larger cap in parallel with each resistor. That'll improve AC stability of the ground voltage a bit.

BTW - that last comment in my previous post should say "...you can leave Q12b in..."
 
Chad,
Right now, i'm doing some measurements

it seams that the banded side (to the left) has about 1.72vdc, and the other side has about 640 mv, aka not what you said it should be

I get a signal at r38, r9 where the components have been removed... but i do not get signal at r7 and r36... normal?

and i only see a few hundred millivolts across r23 and r25, does that mean they are damaged?

Thanks much,

Brad
 
koolscooby said:
Chad,
Right now, i'm doing some measurements

it seams that the banded side (to the left) has about 1.72vdc, and the other side has about 640 mv, aka not what you said it should be
You must be referring to D1, yes? If so, then it sounds like D1 is either damaged or installed backwards. Try putting a new 1N5239B in there and see if that will give you the 9V. What is your supply voltage at? The 'banded side' of D1 should be at +15V, and you should have 9V across D1. Please double-check your supply voltages.

I get a signal at r38, r9 where the components have been removed... but i do not get signal at r7 and r36... normal?
Completely normal. This section of the amp is disabled right now.

and i only see a few hundred millivolts across r23 and r25, does that mean they are damaged?

Thanks much,

Brad
This means there is very little current running through the diff pair. From what you've described above, it sounds like D1 is at fault. If D1 does not give the proper reference voltage, then the current source for the diff pair doesn't work properly. With the front end starved for current, there will not be enough voltage across R23 and R25 to turn on the IRFP044s when they're connected.

To check if R23 and R25 are damaged, turn off the power supply, and measure the resistance of these with the multimeter. It sounds like they might be OK though.
 
ok well ...

first of all, i'm dumb
i had a 10x voltage probe hooked in to my oscope that was set for a 1x, so that screwed my measurements by 10 times

but now thats all fixed

i get +/- 15 v at the PSU

but i only get -8 or from gnd to - plane and about +9 from gnd to + plane, is this normal from the voltage divider?

I have this:

+ -- resistor (1.5k) -- ground plane
- -- resistor (1.5k) -- ground plane

+ -- + plane
- -- - plane

why the heck would it do that? or is that the right setup?, right measurements too?

the d1 is okay, and in correct as per schematic, i put it straight on a 15v ps, and yielded 9.1 drop from 15.1 or so to right around 6.05... so thats fine, put it back in, band to the left

thanks much

brad
 
Umm... I think maybe that's right. It should look like this:
 

Attachments

  • virtual_gnd1.gif
    virtual_gnd1.gif
    1.8 KB · Views: 1,602
Your rail voltages sound too low though. If you have 30V at the power supply, you should get +/-15V on the amp rails. +9/-8V definitely does not sound correct. Perhaps the power supply you're using is loading down when it's connected to the circuit?

At these low voltages, make sure you're using J1b and not J1a.
 
arrgghhhhh :bawling: :smash: :smash:

okay

definately voltage divider isnt working

my psu is only variable up to 25, so i'm going to assume it'll run at 12.5 v, but i'm getting 20v at the + rail and 1.67 v at the - rail... wth? i have it setup as shown in your schematic

why are my output stages on the top of the board turning on? or at least i think they are, they are getting hot... and my poor ***ed attemt at heatsinking is just enough probably not to burn them up :-D does it relate to the voltage divider messing up?


thanks for dealing with my n00bness
 
Ah, I see what's going on here...

OK, short the gate of Q1 to R5 (either end, it doesn't matter). Likewise, short the gate of Q10 to R40. And for good measure, we'll do Q2 and Q11... their gates should be shorted to R6 and R41, respectively. The gates are the pins which were attached to those 221ohm resistors I asked you to remove earlier... also marked G on the PCB silkscreen.

That should take care of your supply voltage problems, and the heating of Q1 and Q10.
 
okay finally!!! :) i have -11.5 on gnd to - and + 12.3v to gnd...

not 15, but thats all the psu puts out :(

it'll work for now i hope until i can get a bigger psu?

well whats next... lets check for 9.1 v across d1...
[banded]12.2v...1.79v[nonbanded]
10.41... good enough 9.1 for me :)

whats next now..

r23 and r25...

r23... -11.5...-7.1
4.4 okay?

r25 -11.6...-7.2
4.4 okay here?

Whats next chad?

Thanks :)
 
OK, so 4.4V is a little over 11mA through each of R23 and R25... a total of about 22mA, which is just a tad low. Now you can adjust VR1 to get a little more current... aim for about 4.6V across each of those resistors.

I have some reservation about a 9.1V zener giving you 10.4V, but we'll look past that for now...

I think we can safely say that your front end is working correctly. The next step is to re-connect and test the power output stages. Unfortunately, you'll need a much bigger power supply to test the outputs, due to the current requirements.

So, in the meantime, I recommend you get your transformers, bridge rectifier, and filter caps set up in preparation for this.
 
awesome...
i have now set 4.6 through them...


powersupply...

with my order of filter caps etc, i'll order more zeners

i lost the ****-small sst505, do you recommend it as a biasing method? or just whatever is used now? (q12a jumpered)
if so i'll try and find the J505 to go in Q12a :) (i hate smt)


i am guessing i should use a Pi filter, as i've read around, so i opened up an old book i found laying around that explained PS some... and read into pi filters etc...

someone (i think bdp) told me that about 2mH is good for the inductors, should they be open core? or ferrite? what do yout hink about this value

filter caps should have what value?
i think i'm going to do a CLC (one Pi) and then a LC, as i've heard its popularity, or should i go CL then CLC?

about how many amps will this amp pull? i'm guessing in the 6 or 7 range? but i have no idea...


thanks much chad, makin progress :)
 
Hehe, I love SMD... you just have to learn it's friendly side. ;) Anyway, the J505/SST505 is really meant more for use with D1b. With D1a, you might as well jumper Q12 and use a resistor for biasing.

Regarding power supply, check out the wiki page here, I've written a lot about the relative merits of various pi filters and what type of inductor to use etc. Right now I'm testing the Hammond 156B choke in a CLC configuration, and this seems to work very well. I'll post some measurements soon.

One channel of standard Aleph-X with 15V rails will pull about 3-5A, depending how heavily you bias it. Mine are set up for 50W into 4ohms from slightly lower rails (around 13V), and bias is set at 5A. Check out the AXE-1 spreadsheet to help you figure out how much bias you'll need for your speakers and the power output that you want.

Something you need to be aware of is the lower voltage you'll get from an L-input filter. Without a cap connected directly to the rectifier bridge, you'll get considerably lower output voltage, while the current load on the transformer and rectifiers will be considerably reduced. This has it's benefits and also some disadvantages. Fortunatley, the Aleph-X with current gain set to 50% is ideally suited for use with an L-input PSU filter. Anyway, read over the power supply section of the wiki, and don't hestitate to post if you have questions. Also, you should download Duncan Amp's PSU Designer II if you don't already have it - it will be indispensible when cooking up your PSU.
 
To get you started with PSUD simulations, here's a simulation file I made for a CLC filter using the Hammond 156B choke. A simulation time of 1000-1500ms is good for this file.

Note that I've used a constant-current 'stepped' load to represent the Aleph-X. The initial startup conditions won't represent real life, but as the simulation approaches 250ms, will give you a glimpse of the no-load operating point of the supply. Keep this in consideration for the voltage ratings you'll need for the main PSU caps. Leave at least a 15% margin above that to account for line surges etc. Once the simulated load steps up to 5A at 250ms, you'll see the voltage drop and stabilize at it's nominal operating point. You can zoom in on the flat section to see the ripple characteristics and get a more accurate idea of it's amplitude.

Note: this simulation file is only for one half of the power supply. Accordingly, you should divide the transformer VA rating in half when doing the secondary voltage calculations.
 

Attachments

  • aleph-x_pi_2.zip
    295 bytes · Views: 873
Things seem a bit slow on this thread recently.Where are all hundreds of boards going?
Anyway I started up myself to convert my aleph 5's to AX.Populated one board with everything but the power transistors and their gate and source resistors.
Tried to check for the ccs and diff pair operation but funny thing happens: they seem to switch on alright but slowly switch off over about 15 seconds.I just could not understand it at all and checked and double checked everything.
Looking at the circuit this morning it dawned on me that maybe what happens is the q3/q8 circuits switch on slowly as the c1 c6 caps charge up slowly then the positive rail gets dragged down ground thru r2/r3/r42/r43 which switches the ccs off.
I guess when you check for diff pair operation you should also leave out r2/r3/r42/r43( or maybe put in r5/r40). Is this correct?
I guess I will check the theory out tonight.
 
protos said:
Things seem a bit slow on this thread recently.Where are all hundreds of boards going?

Right, I've almost finished my amp, waiting for Aluminum pieces to be cutted and milled by a friend of mine (free :) but slow :( ) and then I've to wait some more.

The boards are complete and the next weekend I'll try to switch them on hoping that all will run properly.

I'm building a small case with fans and I have not enough room for big capacitors, I hope that a 40KuF CRC filter will be enough to have a quiet power supply.

As soon as I will have the alu pieces I will post some pics.

Bye!