Preliminary testing of APOX-2
Dale and I did some preliminary testing on the APOX-2
using a simple function generator, and a multimeter
measuring the RMS signal of the output.
Dale's pre-amp seems to be working pretty good.
we feed in between 15 Hz to 100KHz, and the output was pretty flat.
But thats not what we were testing.
We put in approx. 5.25 Vrms across (L+ and L-) into the volume controller, and measured the output at the output connector on the input select board. We then repeated the same test across (R+ and R-)
The data is posted here.
test results
The two outputs seem to track each other very well.
A couple of the resistor values were funny, because Dale's distributor shorted him two resistor values, so he choose the closest he could find.
-Craig Beiferman
Dale and I did some preliminary testing on the APOX-2
using a simple function generator, and a multimeter
measuring the RMS signal of the output.
Dale's pre-amp seems to be working pretty good.
we feed in between 15 Hz to 100KHz, and the output was pretty flat.
But thats not what we were testing.
We put in approx. 5.25 Vrms across (L+ and L-) into the volume controller, and measured the output at the output connector on the input select board. We then repeated the same test across (R+ and R-)
The data is posted here.
test results
The two outputs seem to track each other very well.
A couple of the resistor values were funny, because Dale's distributor shorted him two resistor values, so he choose the closest he could find.
-Craig Beiferman
When I try to open the schematic, my browser gets pushed the URL:
http://www.dipchipelec.com/APOX2/apox1.pdf
Which it tells me cannot be found.
Is it just me?
http://www.dipchipelec.com/APOX2/apox1.pdf
Which it tells me cannot be found.
Is it just me?
Volume - ladder
Only 5k or 10k ladders?
What about 25k, or is this not an issue (with ladders, compared to pot-meters)?
Arne K
Only 5k or 10k ladders?
What about 25k, or is this not an issue (with ladders, compared to pot-meters)?
Arne K
Arne,
Any impedence can be built. Craig just happened to pick those two values for examples.
It all depends on your preamp, whether you put the gain stage before or after the "pot", the output/input impedences of your sources, amplifiers, etc...
As stated above in the thread, if you place the "pot" before the gain stage, 50K or 100K may be more appropriate.
Dale
Any impedence can be built. Craig just happened to pick those two values for examples.
It all depends on your preamp, whether you put the gain stage before or after the "pot", the output/input impedences of your sources, amplifiers, etc...
As stated above in the thread, if you place the "pot" before the gain stage, 50K or 100K may be more appropriate.
Dale
Re Apox1 rev 1 or 2
I think you know what I prefer (version 2). It is beautiful, and i particularly like the relay based SE/Bal thingy which means that you can set up balanced/single ended operation in software per input ... I could definitely NOT have done it better myself -- your ideas are very well executed.
Of course, the pass resistors could be Vishay Bulk Metal Foil -- or Holco -- or Tantalum -- or Carbon composition -- or Wire wound, or Caddock -- or Caddock serpentine -- or SMD ..... I'm getting a bit carried away here, but cost in that position is not really of importance since there are just 4 of them per channel (or 2 if you have a favourite range, or 1 if you are running single ended with a favourite range).
I'm impressed!
The least significant bits will potentially rattle -- have you implemented delay on them so that they don't go into overdrive when changing volume control?
This also looks ideal for a multi-channel volume control. I am about to start on digital crossover for a three way speaker and thus need 6 channels of volume control.
Also, have you guys included provision for setting the sensitity on each channel (so that they each can be set up to be "as loud" regardless of input level? I believe this is called pre-scaling .... which could be implemented as a multiplication of the volume setting ...)
This is the ultimate high-end volume control!
I am still impressed!
Petter
I think you know what I prefer (version 2). It is beautiful, and i particularly like the relay based SE/Bal thingy which means that you can set up balanced/single ended operation in software per input ... I could definitely NOT have done it better myself -- your ideas are very well executed.
Of course, the pass resistors could be Vishay Bulk Metal Foil -- or Holco -- or Tantalum -- or Carbon composition -- or Wire wound, or Caddock -- or Caddock serpentine -- or SMD ..... I'm getting a bit carried away here, but cost in that position is not really of importance since there are just 4 of them per channel (or 2 if you have a favourite range, or 1 if you are running single ended with a favourite range).
I'm impressed!
The least significant bits will potentially rattle -- have you implemented delay on them so that they don't go into overdrive when changing volume control?
This also looks ideal for a multi-channel volume control. I am about to start on digital crossover for a three way speaker and thus need 6 channels of volume control.
Also, have you guys included provision for setting the sensitity on each channel (so that they each can be set up to be "as loud" regardless of input level? I believe this is called pre-scaling .... which could be implemented as a multiplication of the volume setting ...)
This is the ultimate high-end volume control!
I am still impressed!
Petter
Testing Results
Hi,
I finally got the chance to hook everything up last night:
Here was my system:
Home brew DAC (BB PCM1702 4 in parallel per channel) with Zapfilter2 for I/V and output to balanced.
BOSOZ / APOX-2 preamp (5K using Holco H8)
Aleph2 clones
Watt/Puppy clones
This was the first time that I have heard the BOSOZ as well.
Compared to my passive Creek OBH12, there was a huge difference in dynamics, soundstage, and depth of music. There was absolutely no clicking noise when changing volumes with the volume control. (These speakers are pretty efficient, so I expected some pops).
There was no added hum or noise.
When changing inputs, no extra pops/clicks heard.
I am quite pleased!
Hi,
I finally got the chance to hook everything up last night:
Here was my system:
Home brew DAC (BB PCM1702 4 in parallel per channel) with Zapfilter2 for I/V and output to balanced.
BOSOZ / APOX-2 preamp (5K using Holco H8)
Aleph2 clones
Watt/Puppy clones
This was the first time that I have heard the BOSOZ as well.
Compared to my passive Creek OBH12, there was a huge difference in dynamics, soundstage, and depth of music. There was absolutely no clicking noise when changing volumes with the volume control. (These speakers are pretty efficient, so I expected some pops).
There was no added hum or noise.
When changing inputs, no extra pops/clicks heard.
I am quite pleased!
APOX-1 resistor calculations
Hello,
I have attempted to make an excel spreadsheet in an effort
to choose appropriate resistors for the APOX-1
Its now at the bottom of the APOX-1 page.
APOX-1 Resistor Calculations
My hope, is that people will not be confused by my spreadsheet!
Resistors R0-R7 change impedance normally with a binary count, but the bypass switches are giving me nightmares.
Depending on what resistors people choose, I'm not sure
how/when I should switch in the bypass relay? because with the two different bypass settings, an overlap occurs (and is probably desired)
Anyway, I hope this helps 🙂
-Craig Beiferman
Hello,
I have attempted to make an excel spreadsheet in an effort
to choose appropriate resistors for the APOX-1
Its now at the bottom of the APOX-1 page.
APOX-1 Resistor Calculations
My hope, is that people will not be confused by my spreadsheet!
Resistors R0-R7 change impedance normally with a binary count, but the bypass switches are giving me nightmares.
Depending on what resistors people choose, I'm not sure
how/when I should switch in the bypass relay? because with the two different bypass settings, an overlap occurs (and is probably desired)
Anyway, I hope this helps 🙂
-Craig Beiferman
Re: Testing Results
Oh baby! the clones hooked up to the aleph2! How do you like that?
Take a few minutes and make the bosoz an X-bosoz and you'll be even more pleased, the soundstage litereally blows up. 😉
harvardian said:
Oh baby! the clones hooked up to the aleph2! How do you like that?
Take a few minutes and make the bosoz an X-bosoz and you'll be even more pleased, the soundstage litereally blows up. 😉
Apox 1 simulation
Craig,
Your spreadsheet was great! I have looked at this problem before and decided it was a problem, but one I would tackle later. Since you have so kindly done exactly that in Excel, I thought I'd offer some thoughts.
First of all, the way the original spreadsheet is set up is probably not optimal -- the difference between high and low settings is too high. The first thing I did was change the attenuation to -55 and -45dB respectively.
There are other options as well. It is not given that a binary progression is optimal, particuarly given the incredibly high resolution one gets at low volume. It is at high volume settings one wants to assure pretty good resolution. In the end I came up with the one attached. It is not optimal, but probably getting there. Perhaps a higher order exponential progression is what is required ... Hmmmm.
Anyway, I think this is definitely going the right way and welcome any good thoughts on this matter. If a regular logarithmic progression is chosen, it is probably advisable for you to select the ratio between high and low settings, or if you are truly crazy make it programmable ...
So the normal progression will be traversing the high attenuation route until one can cross over to the low res one.
Of course zero volume is handled by the mute relay.
And then it would be opportune to set up the relays so that the pass relay is normally open, while the shunt relays should be normally closed.
Also, the minimum attenuation in the is not shown -- it is zero dB when all relays are open on both ranges. The chart shows this correcty, but the summary segment does not take this into account. In view of this fact, it might be opportune to add the option of a 9th shunt element which is always connected so that the circuit never open circuits. The user can decide whether to implement this option or not.
Petter
Craig,
Your spreadsheet was great! I have looked at this problem before and decided it was a problem, but one I would tackle later. Since you have so kindly done exactly that in Excel, I thought I'd offer some thoughts.
First of all, the way the original spreadsheet is set up is probably not optimal -- the difference between high and low settings is too high. The first thing I did was change the attenuation to -55 and -45dB respectively.
There are other options as well. It is not given that a binary progression is optimal, particuarly given the incredibly high resolution one gets at low volume. It is at high volume settings one wants to assure pretty good resolution. In the end I came up with the one attached. It is not optimal, but probably getting there. Perhaps a higher order exponential progression is what is required ... Hmmmm.
Anyway, I think this is definitely going the right way and welcome any good thoughts on this matter. If a regular logarithmic progression is chosen, it is probably advisable for you to select the ratio between high and low settings, or if you are truly crazy make it programmable ...
So the normal progression will be traversing the high attenuation route until one can cross over to the low res one.
Of course zero volume is handled by the mute relay.
And then it would be opportune to set up the relays so that the pass relay is normally open, while the shunt relays should be normally closed.
Also, the minimum attenuation in the is not shown -- it is zero dB when all relays are open on both ranges. The chart shows this correcty, but the summary segment does not take this into account. In view of this fact, it might be opportune to add the option of a 9th shunt element which is always connected so that the circuit never open circuits. The user can decide whether to implement this option or not.
Petter
Grataku,
It is safe for you to come over now.
I X'd my BOSOZ.
I really needed the 10dB drop in gain.
You were right about the soundstage. I really didn't expect as big of a change.
So, let me know when you/if would like to come over and try out the APOX-2
Best Regards,
Dale
P.S. MikeW, anyone else in the Boston area is also welcome.
It is safe for you to come over now.
I X'd my BOSOZ.
I really needed the 10dB drop in gain.
You were right about the soundstage. I really didn't expect as big of a change.
So, let me know when you/if would like to come over and try out the APOX-2
Best Regards,
Dale
P.S. MikeW, anyone else in the Boston area is also welcome.
APOX-1 Calculations
Peter,
By changing the spreadsheet to -45dB,-55dB
it seems that you killed the high volume attenuation.
(lower left side of graph.)
Which is exactly where you said you wanted the best step size.
(Note: 10dB of attenuation is representing -10dB of gain)
The binary count may be confusing, because the highest value shunt resistor, must be associated with the LSB (Least Significant Bit) of the binary count.
Therefore binary count zero, is actually the loudest volume.
So in order to have decent step size in this range, I needed to lower the gain to -30dB
If I associate the LSB with the lowest value shunt resistor, the binary count does not produce a smooth curve, but rather a complete mess (pattern looks like sin(x)/x function)
Of course, Maybe I confused myself! So please look again to see if your new spreadsheet is what you really wanted.
P.S. I will also add the 9th permanently attached resistor
for people that don't want the circuit to open circuit.
I also don't have a mute relay? Do I need one?
or is the highest attenuation good enough for muting?
(Note: I could also mute by disconnecting the input
on the APOX-IS1)
Thanks,
Craig Beiferman
Peter,
By changing the spreadsheet to -45dB,-55dB
it seems that you killed the high volume attenuation.
(lower left side of graph.)
Which is exactly where you said you wanted the best step size.
(Note: 10dB of attenuation is representing -10dB of gain)
The binary count may be confusing, because the highest value shunt resistor, must be associated with the LSB (Least Significant Bit) of the binary count.
Therefore binary count zero, is actually the loudest volume.
So in order to have decent step size in this range, I needed to lower the gain to -30dB
If I associate the LSB with the lowest value shunt resistor, the binary count does not produce a smooth curve, but rather a complete mess (pattern looks like sin(x)/x function)
Of course, Maybe I confused myself! So please look again to see if your new spreadsheet is what you really wanted.
P.S. I will also add the 9th permanently attached resistor
for people that don't want the circuit to open circuit.
I also don't have a mute relay? Do I need one?
or is the highest attenuation good enough for muting?
(Note: I could also mute by disconnecting the input
on the APOX-IS1)
Thanks,
Craig Beiferman
Coulomb said:
here is Henrik schematic.
http://www.diyaudio.com/forums/atta...hp?postid=67322
Anthony,
Dale confirmed my findings about the soundstage.
I had no doubts, because if I can hear the difference pretty much anyone can.
So a warning to ultra sensitive "golden ears" like Fdegrove: to avoid fainting turn down ear's sensitivity knob by 20 dB. 😉
Dale,
I would be happy to come over on most any weekend of your choice.
Apox 1 settings
Craig,
Please also check out the non-binary sequence of resistors that were used in the spreadsheet I posted. I was actually quite pleased with the resulting response 🙂 The last step will be going from R0 to infinite or the new fixed resistor and is only 2dB.
I have a suggestion though that might kill two birds in one stone.
If you put a relay on each of the pass elements, one can get 4 binary values. If you for example did 0, -10, -20, -30dB in that mode, all that DIY'ers would have to do would be to adhere to this range and the firmware would not need adjustments. This would require another fixed on pass element.
The other comment I have is regarding the sequencing of relays when you get into a mountain range mode (as can easily happen when you mess around with non-binary values), it would be necessary to create a lookup table to ensure correct operation. However that might prove a little complicated.
Muting relays should typically be at the output of a device, and so you might want to consider outputting a muting relay signal (normally off) to provide more tools in the DIY'ers toolbox 🙂
(((New topic: The other functionality which would be great to have would be stackable set -- I need multichannel volume control, but I don't need individual control over each channel -- they all need to track (digital crossover type of application). Would it be possible to set up the addressing so that a number of boards can be used simultaneously?)))
Hope this helps
Petter
Craig,
Please also check out the non-binary sequence of resistors that were used in the spreadsheet I posted. I was actually quite pleased with the resulting response 🙂 The last step will be going from R0 to infinite or the new fixed resistor and is only 2dB.
I have a suggestion though that might kill two birds in one stone.
If you put a relay on each of the pass elements, one can get 4 binary values. If you for example did 0, -10, -20, -30dB in that mode, all that DIY'ers would have to do would be to adhere to this range and the firmware would not need adjustments. This would require another fixed on pass element.
The other comment I have is regarding the sequencing of relays when you get into a mountain range mode (as can easily happen when you mess around with non-binary values), it would be necessary to create a lookup table to ensure correct operation. However that might prove a little complicated.
Muting relays should typically be at the output of a device, and so you might want to consider outputting a muting relay signal (normally off) to provide more tools in the DIY'ers toolbox 🙂
(((New topic: The other functionality which would be great to have would be stackable set -- I need multichannel volume control, but I don't need individual control over each channel -- they all need to track (digital crossover type of application). Would it be possible to set up the addressing so that a number of boards can be used simultaneously?)))
Hope this helps
Petter
Attachments
APOX-1 re-designed
Hello all,
After looking over the APOX-1, I have implemented the excellent suggestions from petter.
The following items have been added to the APOX-1 board.
1.) A permanent attached shunt resistor. to prevent an open
circuit condition, and to easily lower gain with the change of a single resistor.
2.) A relay on the +line to GND for MUTE operation.
(there was already one on the -line, so it will serve double duty, as either the single ended operation control, or the MUTE.
3.) I added another pass resistor and relay. to + and - inputs
for additional range. This will now give us four ranges, each with 255 possible settings. Or 1020 possible positions.
And finally. YES, for digital crossover applications. you can stack these volume control boards. Each APOX-1 board has a fixed address + there are three address offset jumpers for a possiblity of 8 of these boards stacked together.
I will post the version 3 of the schematic tommorow. I am almost done the new layout.
-Craig Beiferman
Hello all,
After looking over the APOX-1, I have implemented the excellent suggestions from petter.
The following items have been added to the APOX-1 board.
1.) A permanent attached shunt resistor. to prevent an open
circuit condition, and to easily lower gain with the change of a single resistor.
2.) A relay on the +line to GND for MUTE operation.
(there was already one on the -line, so it will serve double duty, as either the single ended operation control, or the MUTE.
3.) I added another pass resistor and relay. to + and - inputs
for additional range. This will now give us four ranges, each with 255 possible settings. Or 1020 possible positions.
And finally. YES, for digital crossover applications. you can stack these volume control boards. Each APOX-1 board has a fixed address + there are three address offset jumpers for a possiblity of 8 of these boards stacked together.
I will post the version 3 of the schematic tommorow. I am almost done the new layout.
-Craig Beiferman
WE ARE TAKING ORDERS!
Alright!
The time is here.
All of the prices are posted here
We took a couple of orders already, so now is the best time for
everyone to begin ordering.
Please allow 1 to 6 weeks for delivery as we round up all of the required parts. We will make every effort to expedite your order.
The APOX-IR1, APOX-IS1, APOX-1, and APOX-2 boards will be sent
out on monday for manufacture.
When we get a few more orders will will begin ordering the parts in quantity.
P.S. The new APOX-1 layout looks pretty good.
We will be e-mailing the WIKI group shortly!
Thanks,
-Craig Beiferman
Alright!
The time is here.
All of the prices are posted here
We took a couple of orders already, so now is the best time for
everyone to begin ordering.
Please allow 1 to 6 weeks for delivery as we round up all of the required parts. We will make every effort to expedite your order.
The APOX-IR1, APOX-IS1, APOX-1, and APOX-2 boards will be sent
out on monday for manufacture.
When we get a few more orders will will begin ordering the parts in quantity.
P.S. The new APOX-1 layout looks pretty good.
We will be e-mailing the WIKI group shortly!
Thanks,
-Craig Beiferman
In order to help those who took the jump first, I may try to convince Craig that we will order the parts on speculation. I have FULL confidence that people will really like the kit and that we will have no trouble selling the initial order.
I would hate to have people wait another 5-6 weeks.
Best Regards,
Dale
I would hate to have people wait another 5-6 weeks.
Best Regards,
Dale
- Status
- Not open for further replies.