[BillJGW]

Well, I feel better.

  Thanks, Mandrake.

 

[HOWIE13]

  Well, I feel better.

  Thanks, Mandrake.

Yes, thanks for clarifying that as well, Mandrake. I also heard similar from Frans earlier today.
I do find the low setting useful for some of my poorly digitally remastered analogue classical recordings from the 60-70ties, especially some EMI stuff. With the wide dynamics of classical some of them have been ruined in the remastering process compared to the original LP.
I used to equalise but I find the Polaris BW changes don't affect the surrounding frequencies as equalisers can, so the rest of the sound remains the same, just the harsh top end goes away.

 

 

[ericr]

^^^ Handy feature. I wonder if it would get me listening to some of my poorly recorded music again.

 

[iancraig10]

For info regarding the bandwidth settings. The widest bandwidth setting is (in effect) 'off'.

Here are the effects of using it:






Green - high
Orange - mid
Red - low

This is Rev 1.1 (30 ohms output impedance)

High settings take the FR way beyond the threshold of hearing.

Not to mention the output impedance settings as well.

 

[mandrake50]

Graph is for which configuration, the original from configuration from Jeremy, or (newer) changed per Frans' thoughts?

 

[HOWIE13]

Graph is for which configuration, the original from configuration from Jeremy, or (newer) changed per Frans' thoughts?

Probably Jeremy's original as I don't think the Mid setting falls off as far with Frans' more recent changes.
It would be good to compare the two graphs as I find it easier to conceptualise that way.
However, Ian's graph does make the point that this High setting would not interfere with the audible native sound which is most reassuring.

 

[Soundsgoodtome]

Wait, so I have one of the first units built and probably have v1.

So the differences are longer wait time on the relay and less drop off on the bandwidth switch between the selections? Do I have that right?

 

[iancraig10]

Wait, so I have one of the first units built and probably have v1.

So the differences are longer wait time on the relay and less drop off on the bandwidth switch between the selections? Do I have that right?

Very few v1 boards are out. The graph is for v1.1 with (I think) 30 ohms output impedance. V1.1 is the one that ships now and is the one that I'll be listening to next week.

You know, you have me thinking. I thought switch on time had been lowered to something like 2 seconds on v1.1 and was slightly longer on v1. Does the board have a marking on it?

If you check DIYAH, Frans has put a full explanation there. We're not talking anything massive in changes here with the BW settings.

Someone earlier posted a really relevant post where they referenced an earlier message from Frans that details the 1.1 changes and I thought he had lowered the time. Now I'll be timing it for two seconds!!!

 

[micmacmo]

I think there are actually three versions of the Project Polaris. The rare v1.0, which is the version that went to some early adopters. There's version v1.1, which was the recent production model (which is graphed in iancraig10's post above). And now a new version. Fran's explains the differences, including an additional graph of the FR of the new version, in this post on DIYAH.
 
EDIT:
BTW, my board is marked PROJECT POLARIS R1.0 by the standoff beside the input RCA jacks. My turn-on time is about 8 seconds. I'm satisfied with my amp, regardless of version #.

 

[iancraig10]

Yes, that third one is less rolled down and so will sound slightly more aggressive on the medium and lower bandwidth settings but the very same on high, which is what most seem to use in any case.

Below 1kHz is exactly the same . It's not released yet though.

 

[HOWIE13]

Very few v1 boards are out. The graph is for v1.1 with (I think) 30 ohms output impedance. V1.1 is the one that ships now and is the one that I'll be listening to next week.

You know, you have me thinking. I thought switch on time had been lowered to something like 2 seconds on v1.1 and was slightly longer on v1. Does the board have a marking on it?

If you check DIYAH, Frans has put a full explanation there. We're not talking anything massive in changes here with the BW settings.

Someone earlier posted a really relevant post where they referenced an earlier message from Frans that details the 1.1 changes and I thought he had lowered the time. Now I'll be timing it for two seconds!!!

I have v1.1 and the switch on time is 7 seconds. That's fine for me.  

 

[iancraig10]

Oh ok. Then maybe I read it where Frans was suggesting changing some capacitor values somewhere on DIYAH.

I think he mentions 2 seconds somewhere. Comparatively quite a lot, but not a lot in reality!! I'll time mine when it arrives.

Aha ... Found it ......





I changed a few values in Polaris:
The first change has to do with the start-up relay which is now faster.
This can be faster as there is no tube that needs to warm up.
For this change (the start up delay becomes around 2 seconds) 2 bleeder resistors (R5) need to be lowered in value.
This is possible because the output stage has short-circuit protection limiting the current.
so.... R5 -> 33 Ohm (was 330 Ohm)

To make the relay engage faster the timing is altered by changing R4.
So .... R4 -> 56k (was 150k)

The difficult part will be mounting the resistors after you removed the old ones.
The reason for this is the thick PCB that makes it very hard to remove the solder inside the through hole metallization in the pads that are connected to the ground plane.
ONLY remove these resistors (R4 and R5's) if you are experienced in de-soldering.
Those with a 0.7mm drill can also drill the pad open again (is what I do in such cases)

As these parts aren't in the audio path one can also mount 'non-audiophile' resistors but as it is open frame similar DALE type resistors may look nicer.

BUT there is another way...
As the values become smaller you can also mount small resistors in parallel on the bottom of the PCB (is what I did) and you can always paint them black with a dab of paint if you want to conceal them.

R4 needs 82k in parallel to the 150k that is already in there. This will give 53k which is just as good as 56k in this case.
Both R5's need a 39 Ohm resistor in parallel to the 330 Ohm that is already in there. This will give 35 Ohm which is just as good as 33 in this case.

as a result the 'waiting time' is now around 2 seconds. Only perform this mod if you feel it is a worthwhile improvement.

Another thing I changed is the bandwidth setting.
Jeremy wanted to use the bandwidth to lower treble on treble happy headphones and thus made the low and medium bandwidth setting a bit too much (IMO)
In the Low setting (where C10 is 820pF) the upper frequency (-3dB) is 12kHz. The medium bandwidth (-3dB) is around 17kHz

My intention (as in the CHAmp design) is to have 3 bandwidth settings that can be selected but do not interfere with the sound that much.

The high BW setting is for those that do not want any phase shifts nor roll off in the audible band and far outside of it.
Perfect for analog sources and DAC's with 'fast' or 'steep' digital filter settings and have proper HF audio filtering

The mid BW setting is for those that may have (oversampling) NOS DAC's or do not feel they need 300kHz bandwidth.
High bandwidths can (in some rare cases) have a negative influence on the sound when aliasing with HF garbage is present.
In this setting there is no roll-off in the treble (flat within 0.5dB) up to 20khz and the (very gentle 6dB/oct) slope is -3dB at 35khz which is wider than most headphones and than most recordings actually have.
A recommended setting when using NOS DAC's.

The low BW setting is now set at 20kHz (-3dB) and has a very similar roll-off as NOS DAC's show at 44.1/48kHz files.
Some like the gentle roll-off those DAC's have and the Polaris emulates (well almost) this behaviour when using 'faster' DAC's
It will not remove brightness from headphones but can add some slight warmth to music while retaining frequency extension.

To create the changes to the values/effects I wrote above 4 capacitors have to be replaced.
As these caps have 1 pad grounded a similar level of difficulty (in desoldering) occurs as above.
Alas the values have to become smaller and thus paralleling is not possible.

If you want these changes then the options are to desolder the caps (C10's and RC11's) next to the BW jumpers and remove the solder (or drill out the holes with 0.7mm drill).
Another option is to cut-away the caps (and thus leave the pins in the PCB).
You can then solder the new caps on the top side OR bottom side of the PCB on the pads.

C10's -> 470pF (was 820pF)
C11's -> 270pF (was 330pF)
In case of C11 (the difference with 270pF isn't that big) you can opt to only replace the 820pF C10 and leave C11 as it was.

Again these mods are only needed IF you think the BW settings are limiting too audible.
For those having some bright headphones and use the low BW setting to tame the treble (is what Jeremy wanted) you can skip this mod.

This is only needed for Polaris amps sold before Sept 2014.
Later amps are probably already shipping with the updated values (BW and start-up time) and IF you feel it is needed.

Also NOTE:
The manual dated aug 3 on the download page has INCORRECT values for C10, C11, R4 and R5.
Please check for the latest manual date

Read more: http://diyah.boards.net/thread/681/project-polaris#ixzz3EpoMXUTd
26th August from Solderdude.

 

[mandrake50]

I had links to the thread a few posts ago... FWIW

 

[Soundsgoodtome]

Unless the serial numbers for the v1.0 is single digit only, I believe to have one. This was built with the first batches by Jeremy.
 
Quote:

Look what just arrived! Looking forward to the HE4 and HE560 sessions soon! Going to give it 3 days of run time before any impressions.

 

[BillJGW]

This is all great info!  I 'm a visual person so it's great to see all of this presented in graphs.  Thanks Ian and micmacmo (and solderdude on diyah).  Two big thumbs up!