Re: Current limiting is preventing it coming up correctly
I look at your question and see references to resistor and transistor
numbers. I don't keep a copy of the schematic around (after all, it's
not mine) and I know I can burrow through a zillion posts looking for
it, or hope that someone who actually keeps a copy handy is smart
and motivated enough to respond.
Jason Hubbard said:
I look at your question and see references to resistor and transistor
numbers. I don't keep a copy of the schematic around (after all, it's
not mine) and I know I can burrow through a zillion posts looking for
it, or hope that someone who actually keeps a copy handy is smart
and motivated enough to respond.
I have been reading your posts and as a successful builder of the Aleph-X I have been unwilling to respond for a couple of reasons.
First, there is the lack of context as mentioned by Nelson, and second you have used the J500 style modified CCS which is an obscure variant that few have built AFAIK. I for one would not feel confident in commenting. So I'm not surprised that no one has spoken up.
I think you're on the right track in creating and using a "real" power supply rather than your bench supply.
IME if the voltage across a CCS zener is less than the zener voltage it means that your rail voltage isn't high enough.
Graeme
First, there is the lack of context as mentioned by Nelson, and second you have used the J500 style modified CCS which is an obscure variant that few have built AFAIK. I for one would not feel confident in commenting. So I'm not surprised that no one has spoken up.
I think you're on the right track in creating and using a "real" power supply rather than your bench supply.
IME if the voltage across a CCS zener is less than the zener voltage it means that your rail voltage isn't high enough.
Graeme
Thanks for a response guys
With all the boards that were made i figured there would still be a number of amplifier builds still in progress and consequently a little more activity within this thread.
This has been an uphill struggle, though not one without its rewards - i now understand far more about the circuit than i ever did (though i am clearly no expert).
Nelson, I have used the v1.0 boards that were made in abundance some years ago and any part numbering is as per the diagram that accompanies that (not greys original , but with minor modifications to accomodate ideas brought into the mix by other forum members). I appreciate your comments but as a lone DIYer I was not expecting you to reply!
Graeme - There was a UK group buy for boutique parts which i was included in - it used the J505 constant current diode (though i have swapped it out for a more standard 1k5 resistor for R17 and no J505 at all, there was no difference in the low voltage across the zener so i put the J505 back in). Rails are +/- 15v as per the "stock" design so zener voltage should be as expected. I've varied the rails by a volt or so without any change in Zener voltage. Will experiment some more, just anxious not to smoke anything!
Wolfpako - Happy to talk, mail me.
With all the boards that were made i figured there would still be a number of amplifier builds still in progress and consequently a little more activity within this thread.
This has been an uphill struggle, though not one without its rewards - i now understand far more about the circuit than i ever did (though i am clearly no expert).
Nelson, I have used the v1.0 boards that were made in abundance some years ago and any part numbering is as per the diagram that accompanies that (not greys original , but with minor modifications to accomodate ideas brought into the mix by other forum members). I appreciate your comments but as a lone DIYer I was not expecting you to reply!
Graeme - There was a UK group buy for boutique parts which i was included in - it used the J505 constant current diode (though i have swapped it out for a more standard 1k5 resistor for R17 and no J505 at all, there was no difference in the low voltage across the zener so i put the J505 back in). Rails are +/- 15v as per the "stock" design so zener voltage should be as expected. I've varied the rails by a volt or so without any change in Zener voltage. Will experiment some more, just anxious not to smoke anything!
Wolfpako - Happy to talk, mail me.
Re: Thanks for a response guys
I once wrote a note
And placed it in a bottle
I knew its words by rote
But since then, years have passed a-lotta
(Just made that one up and don't have time to perfect it...hopefully you catch my drift.)
Me? I've got two problems.
1) Ages have passed and while I could once quote the part numbers off the top of my head, it's all gone now. The fact that scads of others had their way with the tender young thing, changing part numbers (and in fact the parts themselves) and I quickly gave up any attempt to keep track of variations.
2) Diagnosing problems at a distance drives me absolutely bonkers. You'll not find many fix-it threads that I've participated in. Others seem to delight in such things. My hat's off to them.
Grey
Jason Hubbard said:
I once wrote a note
And placed it in a bottle
I knew its words by rote
But since then, years have passed a-lotta
(Just made that one up and don't have time to perfect it...hopefully you catch my drift.)
Me? I've got two problems.
1) Ages have passed and while I could once quote the part numbers off the top of my head, it's all gone now. The fact that scads of others had their way with the tender young thing, changing part numbers (and in fact the parts themselves) and I quickly gave up any attempt to keep track of variations.
2) Diagnosing problems at a distance drives me absolutely bonkers. You'll not find many fix-it threads that I've participated in. Others seem to delight in such things. My hat's off to them.
Grey
Hello Jason,
As you can see from my posts above i'm no expert but I'll try to help. If I where you i'd start by doing three things:
First, go back to grey's original schematic and implement the ccs as per the schematic (ie zener and 1k5 for bias).
Second, change R11 and 33 back to the original value of 47k5
Third, build a psu capable of delivering at least 225 VA per channel(trannie like 2x15V @7,5A and then some capacitance 10mf/.22R 10w/10mf on both lines for the test).
Now from what i've read above you've got something like 4.3V across R23 and 25 (just check that they don't differ to much). This seems about right and suggests the ccs is working ok. I wouldn't be to worried about the voltage reading across the zener. 0.3v (9.1-8.8) is less then 5% deviation and might well be caused by error in your meter. Thus stick with Grey's original schematic since this is known to work.
Now once you've done the above, ground the inputs and remove the load across the outputs. Grounding the inputs is recommended for the pumpkin (similar circuit) and i believe it was also recommended in the wiki. Anyway it helps during setup ime. Make shure that V1 and V3 are at the same setting.
power up the unit, you haven't broken anything so far so I guess you'll be ok. Give the circuit time, it needs to warm up. Just watch for smoke or weird sounds. Keep monitoring the voltage across one of the source resistors and the absolute dc offset (output to ground).
After about ten minutes the absolute dc offset should be lowering slowly, be aware it will still wander around. after about 20 minutes you should be able to start adjusting the dc offset with V2. Try and set it at +/- 1V. Wait ten minutes and adjust it again.
If the absolute dc offset remains stable at around +/- 1V, check if both sides have equal voltages across the source resistors. This you can adjust with V1 and V3 (you seem to be able to adjust the current, can you do this per channel? if not check the pinouts of Q3,8,4 and 9). Be aware that adjusting one will affect the other. You should get something like 0.4 to 0.5V across each resistor if i remember correctly from the wiki (3.4A current draw total)
Now check the differential offset (across the outputs). if it is less then .5 volts power down the unit and reattach the 7R load across the outputs. Repeat the steps above.
Now remove the short to ground on the input and plug in a source (one that may be destroyed and check it is ac coupled and not giving any dc on the output, a walk or discman can do the trick). repeat the steps above.
Now if the absolute offset is lower then +/- 2V and the differential offsett less then .5V, power down and attach a cheap test speaker that may be destroyed. Power up the unit, wait 20 minutes while monitoring voltages and then press play.
Hope this helps.
regards
Joris
Edit: forgot to mention that only after warm up the voltage across the source resistors should be equal to some degree.
As you can see from my posts above i'm no expert but I'll try to help. If I where you i'd start by doing three things:
First, go back to grey's original schematic and implement the ccs as per the schematic (ie zener and 1k5 for bias).
Second, change R11 and 33 back to the original value of 47k5
Third, build a psu capable of delivering at least 225 VA per channel(trannie like 2x15V @7,5A and then some capacitance 10mf/.22R 10w/10mf on both lines for the test).
Now from what i've read above you've got something like 4.3V across R23 and 25 (just check that they don't differ to much). This seems about right and suggests the ccs is working ok. I wouldn't be to worried about the voltage reading across the zener. 0.3v (9.1-8.8) is less then 5% deviation and might well be caused by error in your meter. Thus stick with Grey's original schematic since this is known to work.
Now once you've done the above, ground the inputs and remove the load across the outputs. Grounding the inputs is recommended for the pumpkin (similar circuit) and i believe it was also recommended in the wiki. Anyway it helps during setup ime. Make shure that V1 and V3 are at the same setting.
power up the unit, you haven't broken anything so far so I guess you'll be ok. Give the circuit time, it needs to warm up. Just watch for smoke or weird sounds. Keep monitoring the voltage across one of the source resistors and the absolute dc offset (output to ground).
After about ten minutes the absolute dc offset should be lowering slowly, be aware it will still wander around. after about 20 minutes you should be able to start adjusting the dc offset with V2. Try and set it at +/- 1V. Wait ten minutes and adjust it again.
If the absolute dc offset remains stable at around +/- 1V, check if both sides have equal voltages across the source resistors. This you can adjust with V1 and V3 (you seem to be able to adjust the current, can you do this per channel? if not check the pinouts of Q3,8,4 and 9). Be aware that adjusting one will affect the other. You should get something like 0.4 to 0.5V across each resistor if i remember correctly from the wiki (3.4A current draw total)
Now check the differential offset (across the outputs). if it is less then .5 volts power down the unit and reattach the 7R load across the outputs. Repeat the steps above.
Now remove the short to ground on the input and plug in a source (one that may be destroyed and check it is ac coupled and not giving any dc on the output, a walk or discman can do the trick). repeat the steps above.
Now if the absolute offset is lower then +/- 2V and the differential offsett less then .5V, power down and attach a cheap test speaker that may be destroyed. Power up the unit, wait 20 minutes while monitoring voltages and then press play.
Hope this helps.
regards
Joris
Edit: forgot to mention that only after warm up the voltage across the source resistors should be equal to some degree.
Think i'm almost there
Joris - Many thanks for your help.
Ok - i have built a power supply and that has made a big difference (i can get both sides of the circuit to come up now).
I changed R11 & R33 back to 47.5K & grounded the inputs as suggested. For the moment at least i have left the CCS as is (it does seem to be working correctly).
I had 0.5v across all the source resistors but struggled with abs DC offset (unable to get it below 1.5volts) - changing R24 from 470 to 330 ohms gave me the adjustment range i needed to zero that out, but i still have relative DC offset of about 140mv that i'm trying to "tweak out".
If i adjust Vr1/Vr3 to give zero relative voltage then the voltage across the source resistors is no longer equal (about 60mv difference left to right).
At the moment i have the voltage across each source resistor equal at 0.54v, that seems to be the "sweet spot" for everything balancing.
Should i reduce relative DC offset & have slightly different voltages across the source resistors or leave with equal voltages but with 140mv relative offset? Or some other solution?
Joris - Many thanks for your help.
Ok - i have built a power supply and that has made a big difference (i can get both sides of the circuit to come up now).
I changed R11 & R33 back to 47.5K & grounded the inputs as suggested. For the moment at least i have left the CCS as is (it does seem to be working correctly).
I had 0.5v across all the source resistors but struggled with abs DC offset (unable to get it below 1.5volts) - changing R24 from 470 to 330 ohms gave me the adjustment range i needed to zero that out, but i still have relative DC offset of about 140mv that i'm trying to "tweak out".
If i adjust Vr1/Vr3 to give zero relative voltage then the voltage across the source resistors is no longer equal (about 60mv difference left to right).
At the moment i have the voltage across each source resistor equal at 0.54v, that seems to be the "sweet spot" for everything balancing.
Should i reduce relative DC offset & have slightly different voltages across the source resistors or leave with equal voltages but with 140mv relative offset? Or some other solution?
Hello Jason,
Nice to hear you're doing ok. You are right about the ccs, if it works let it be.
I would set V1/3 to equal current, so leave it like you have it now. The differences in voltage drop over the source resistors will in part be caused by small variations in both the source resistors and the power mosfets. You could match the source resistors and power mosfets or just swap them around a bit and see what gives the best result for the relative offset.
The relative dc offset that remains is probably due to components not being matched close enough. If you want to further weed out the relative dc offset look at better matching. From what I've read the input fets are the most critical and Q2/11 should be matched as well.
Edit: do you have the input fet's thermally coupled? If not do so but keep them electrically isolated.
The 140mv relative dc offset is a bit high but not high enough to be a serious worry. I would just finish up the test setup for both channels and start listening to some music. You can always tweak and perfect the unit later.
regards,
Joris
Nice to hear you're doing ok. You are right about the ccs, if it works let it be.
I would set V1/3 to equal current, so leave it like you have it now. The differences in voltage drop over the source resistors will in part be caused by small variations in both the source resistors and the power mosfets. You could match the source resistors and power mosfets or just swap them around a bit and see what gives the best result for the relative offset.
The relative dc offset that remains is probably due to components not being matched close enough. If you want to further weed out the relative dc offset look at better matching. From what I've read the input fets are the most critical and Q2/11 should be matched as well.
Edit: do you have the input fet's thermally coupled? If not do so but keep them electrically isolated.
The 140mv relative dc offset is a bit high but not high enough to be a serious worry. I would just finish up the test setup for both channels and start listening to some music. You can always tweak and perfect the unit later.
regards,
Joris
220uF bypass
Hi,
I've seen a photo of the aleph x 1.0 pcb with the 220uF's bypassed with wima 0.01uF caps. the 220uF's leads are
raised up to fit the wima's. Does this bypass help a lot
or not necessary?
also, 25V caps are fine here or they must 35V?
I have the boards and building the standard 40W version.
thx
Hi,
I've seen a photo of the aleph x 1.0 pcb with the 220uF's bypassed with wima 0.01uF caps. the 220uF's leads are
raised up to fit the wima's. Does this bypass help a lot
or not necessary?
also, 25V caps are fine here or they must 35V?
I have the boards and building the standard 40W version.
thx
Hi William,
how big were your bypass caps that you mounted on
the bottom side of the pcb. I got the wima .01 63V
ones that are pretty thick.
I thought of mounting on bottom as well but the width
seems to be too thick where it would not allow the FET's
to mount flush to the heatsink. maybe I need to bend
the FET leads more?
thanks for replying
how big were your bypass caps that you mounted on
the bottom side of the pcb. I got the wima .01 63V
ones that are pretty thick.
I thought of mounting on bottom as well but the width
seems to be too thick where it would not allow the FET's
to mount flush to the heatsink. maybe I need to bend
the FET leads more?
thanks for replying
input coupling capacitor
Can anyone suggest a suitable input coupling capacitor - was thinking of something in the order of 6.8uf - would that be OK or will i suffer premature frequency rolloff?
Also - i am feeding 0.5v p-p 1khz square wave in (via 4.7uf capacitors) and seeing 2v p-p coming out - is that about the right sort of gain to expect? (it is a perfect square wave coming out, no rounded edges or overshoot)
Can anyone suggest a suitable input coupling capacitor - was thinking of something in the order of 6.8uf - would that be OK or will i suffer premature frequency rolloff?
Also - i am feeding 0.5v p-p 1khz square wave in (via 4.7uf capacitors) and seeing 2v p-p coming out - is that about the right sort of gain to expect? (it is a perfect square wave coming out, no rounded edges or overshoot)
Aleph-X gain
Hi William,
Thanks for the response - input resistors R18, 19, 28, 29 are all 10K - not sure what impedence that would be ( i know r28 & 29 are 68k in the original schematic from grey, giving an input impedence of about 47k).
I don't have my notes with me at the moment regarding the resistors that set the gain (i'm working away from home on my company laptop and don't have access to my personal data), but they should be as per the schematic associated with the hifizen/grataku rev 1.0 PCB's (sorry if that's not much help).
I will post details of the resistor values when i get home on Thursday.
Jason
Hi William,
Thanks for the response - input resistors R18, 19, 28, 29 are all 10K - not sure what impedence that would be ( i know r28 & 29 are 68k in the original schematic from grey, giving an input impedence of about 47k).
I don't have my notes with me at the moment regarding the resistors that set the gain (i'm working away from home on my company laptop and don't have access to my personal data), but they should be as per the schematic associated with the hifizen/grataku rev 1.0 PCB's (sorry if that's not much help).
I will post details of the resistor values when i get home on Thursday.
Jason
single ended
Hi,
how many have built their aleph X and use it with single ended source?
my reading came upon this issue:
{I can adjust the bias nicely and can also adjust the relative DC offset to zero. BUT as soon as I connect the negative XLR input to ground (For RCA input) I get a relative DC input of 4V! Absolute
DC offset is 6V with XLR and 8V!!! negative XLR input grounded (for RCA input). What did I wrong?}
was there a resolve for this problem?
is there an audio performance toll for not using a balanced source?
thanks.
Hi,
how many have built their aleph X and use it with single ended source?
my reading came upon this issue:
{I can adjust the bias nicely and can also adjust the relative DC offset to zero. BUT as soon as I connect the negative XLR input to ground (For RCA input) I get a relative DC input of 4V! Absolute
DC offset is 6V with XLR and 8V!!! negative XLR input grounded (for RCA input). What did I wrong?}
was there a resolve for this problem?
is there an audio performance toll for not using a balanced source?
thanks.
