Hi BV,
Thanks. That does seem to run a little better. I have a few questions if I may.
1) It seems to be very low gain. 3.5V input doesn't bring it to clipping. I changed R24 to 20K and now it will clip at 1.8V input. If this the correct way to increase gain?
2) I see you changed Q107 and Q108 to the TO-220 package. Some folks have indicated those are not enough to drive 5 pair of outputs. What reason did you change those?
3) You removed C107 & C109. Why?
Can someone who know how compare this to the Wolverine? It would be interesting to know how they compare.
Thanks, Terry
Thanks. That does seem to run a little better. I have a few questions if I may.
1) It seems to be very low gain. 3.5V input doesn't bring it to clipping. I changed R24 to 20K and now it will clip at 1.8V input. If this the correct way to increase gain?
2) I see you changed Q107 and Q108 to the TO-220 package. Some folks have indicated those are not enough to drive 5 pair of outputs. What reason did you change those?
3) You removed C107 & C109. Why?
Can someone who know how compare this to the Wolverine? It would be interesting to know how they compare.
Thanks, Terry
Man you get your boards quickly. I ordered mine on 5-13 and they still haven't reached the States. Are those the CFA-XH 'IPS boards?
Any chance you will be creating Gerbers for any of the other IPS?
Thanks, Terry
Imbalance in VAS loading
Hello,
I´ve been playing around with LTspice lately. Especially with the CFA-X model.
I have made an interesting observation by loading the vas unevenly.
Based on the simulation uneven loading at some resistor values actually seems to lower the total harmonic distortion while H2 is increased.
I am thinking of installing a pot together with the load resistors for listening experiments.
I just need to take the thumb out of the **** and finish the build during weekend.
BR,
J
Hello,
I´ve been playing around with LTspice lately. Especially with the CFA-X model.
I have made an interesting observation by loading the vas unevenly.
Based on the simulation uneven loading at some resistor values actually seems to lower the total harmonic distortion while H2 is increased.
I am thinking of installing a pot together with the load resistors for listening experiments.
I just need to take the thumb out of the **** and finish the build during weekend.
BR,
J
Attachments
OK, this is the spice directive I used for both.
;ac oct 10 10 20000000
.options plotwinsize=0
.options numdgt=8
.param freq 10k
.param num_fft_pts 4096
.param timestep {1/{{freq} * {num_fft_pts}}}
.four {freq} 19 8 V(Vout)
.tran 0 {15/freq} {2/freq} {timestep}
Edit; I am adding a second chart. Both amps set to 130v p-p output. +-80vdc rails.
Attachments
Last edited:
Who did you order from? I used my usual supplier, Sitopway, and they are quick. The usually ship by DHL, so typical transit time is only three days.
Those are the CFA-XH v1.3 boards.
I will ultimately do some different IPS but since I'm not independently wealthy they will happen over time. It is a question of which ones and which first. I'm thinking either 'Spooky Leach' or 'Wolverine' next. I also like the idea of scaling the SymAsym or doing a different 'Blameless' and cooking up one of my own. Can't let OS have ALL the fun...
Hi Jason,
I'll try Sitopway next time.
It doesn't cost anything to make Gerbers,
It can cost a surprising amount of time, at least for me anyhow.
The design doesn't just flow out of my head and into a tangible form as easy as folks like alex mm seem to be able to do it, even when someone else like OS has made the basic framework. I use different spacings so sometimes the layout, though functionally similar can be aesthetically different. I'm also picky about how things are placed / spaced.
I also don't make any bones about utilizing the top copper since I am not catering to the home brewed PCB crowd either. I very much with you on the superiority of using a board house for the purpose of making the PCB.
I'm open to suggestions 'by popular demand' on which IPS should be next in line for massaging.
Sorry, I didn't mean to imply it was easy. I have been trying to learn it and can't even get the components placed on the board. I really, really appreciate all you do for us here.
My vote goes for the Symasui. I will be getting a laser printer later this month so I will probably try etching some of these. I'm not sure how many other people are going to want to build these and I always order at least 10 boards and I'm sure you have to do that too. I think I will eventually use all my OPS boards but not likely 5 pair of the same IPS. Maybe etching is the best option.
Blessings, Terry
My vote goes for the Symasui. I will be getting a laser printer later this month so I will probably try etching some of these. I'm not sure how many other people are going to want to build these and I always order at least 10 boards and I'm sure you have to do that too. I think I will eventually use all my OPS boards but not likely 5 pair of the same IPS. Maybe etching is the best option.
Blessings, Terry
With the whirlwind that OS can be at times I'm not even 100% sure exactly how many IPS there are. I recall distinctly these:
Wolverine - Next generation Badger, blameless type VFA
Gnome - NAD based CFA (I still thought the moniker GoNAD was funny, I liked it)
Symasui - SymAsym / Sansui based VFA
Spooky Leach - Leach based mirror image VFA
CFA-XH - VSSA based with Hawksford Cascode CFA
Did I miss any? At one point OS suggested trying to keep 'official' offerings to four for the reason of supporting builds better with a limited number of IPSs.
It is the paradox of choice; the more choices are available the less likely a decision will be made. It happens in retail all the time. Offer a consumer two or three choices and they will likely buy on of them, but offer them 20 or 30 choices and they often leave without a purchase.
I'm inclined to eventually do most of these listed, mainly because they are distinct from each other and have at the very least some historical / sentimental significance. OS also mentioned cooking up more when he is settled in TN, oh my
!
Wolverine - Next generation Badger, blameless type VFA
Gnome - NAD based CFA (I still thought the moniker GoNAD was funny, I liked it)
Symasui - SymAsym / Sansui based VFA
Spooky Leach - Leach based mirror image VFA
CFA-XH - VSSA based with Hawksford Cascode CFA
Did I miss any? At one point OS suggested trying to keep 'official' offerings to four for the reason of supporting builds better with a limited number of IPSs.
It is the paradox of choice; the more choices are available the less likely a decision will be made. It happens in retail all the time. Offer a consumer two or three choices and they will likely buy on of them, but offer them 20 or 30 choices and they often leave without a purchase.
I'm inclined to eventually do most of these listed, mainly because they are distinct from each other and have at the very least some historical / sentimental significance. OS also mentioned cooking up more when he is settled in TN, oh my
!
No problem, I wasn't suggesting any offence.
As for a layout program that is pretty easy to use, I like Abacom's Sprint Layout. It is not free, but it is one of the easier ones to use and it isn't a crazy amount of money to buy. I tried Eagle but ultimately it just gave me a headache and it put me off trying many others.
Personally, I don't mind gradually doing the IPSs and ordering some as they are ready. It just means if someone is holding out for the last one I end up doing they may have to wait for a while.
Thimios beat us to the punch on most of these. His findings were the same as I usually find...All good amps sound good. The last two days I spent some time listening to some of my older amps. Both Krell KSA50 clones, Leach Low TIM and ESP P101. I hadn't used them in a long time and thought I might salvage one or more to use the PSUs and heatsinks for new builds. Forget it. They all sound lovely, VERY lovely! I will be salvaging the Leach Superamp and Aleph-X but the rest will live on.
Let me know which IPS Gerbers you plan to do next so I don't waste time etching them. Hopefully I will have my CFA-XH boards soon and that will give me something to work on.
Blessings, Terry
I will be salvaging the Leach Superamp and Aleph-X but the rest will live on.Let me know which IPS Gerbers you plan to do next so I don't waste time etching them. Hopefully I will have my CFA-XH boards soon and that will give me something to work on.
Blessings, Terry
Anyone ever use Express pcb, it is a free software tool for doing board layout? I haven't tried to do any of this yet but it looks interesting to try.
Jason,
How much different is the Gonad from the CFA-XH, aren't they similar in many ways? Did Thimios build both and compare them subjectively for sound?
Jason,
How much different is the Gonad from the CFA-XH, aren't they similar in many ways? Did Thimios build both and compare them subjectively for sound?
Express PCB is linked to that particular board house, IIRC. The software is free, but the boards are comparatively expensive. I prefer to use a tool that generates industry standard files so I choose my supplier. Some board houses can accept project files from different software packages and they generate the gerbers and drill files themselves.
I will have to put both schematics in front of me and look, but I believe they are distinctly different. They just happen to use low impedance feedback nodes and as a result are so-called CFAs. Thimios built a pile of stuff, so I'm not too sure what he has compared to what. Send him a PM maybe and ask?
not gone yet ...
My "downtime" will most likely be 6/7 to 7/7 2014.
There are only a few topologies/variants that I have not touched on
that would go well with the slew EF3.
1. IC based .... either a high Vcc based LMxxxx based one, or a level shifted
symmetrical crown/QSC IPS with 15V supplies/TL071-5532 IC.
2. A simple ASKA/DX bootstrapped LIN amp without any CCS's.
Just simple resistors and a cascoded (low voltage)/ regulated LTP.
The regulated lower voltage would preserve PSRR without CCS's !
3. The "superpair" KODA type amp (below). Apex (audio) uses the
symmetrical version and Koda ... just a single ended version.
There are others ... but most fall into the "variant" category.
A fet IP pair "john curl" type IPS is the only other unique one.
Of course , with an error correcting or HEXFET OPS , there are
even more options .
PS- almost forgot .... a "wolverine" type input stage running a
Hawksford with a level shifter (like the original Luxman 120a -below 2) ...
might also make for listening pleasure.
OS
With the whirlwind that OS can be at times I'm not even 100% sure exactly how many IPS there are. I recall distinctly these:
Wolverine - Next generation Badger, blameless type VFA
Gnome - NAD based CFA (I still thought the moniker GoNAD was funny, I liked it)
Symasui - SymAsym / Sansui based VFA
Spooky Leach - Leach based mirror image VFA
CFA-XH - VSSA based with Hawksford Cascode CFA
Did I miss any? At one point OS suggested trying to keep 'official' offerings to four for the reason of supporting builds better with a limited number of IPSs.
It is the paradox of choice; the more choices are available the less likely a decision will be made. It happens in retail all the time. Offer a consumer two or three choices and they will likely buy on of them, but offer them 20 or 30 choices and they often leave without a purchase.
I'm inclined to eventually do most of these listed, mainly because they are distinct from each other and have at the very least some historical / sentimental significance. OS also mentioned cooking up more when he is settled in TN, oh my
My "downtime" will most likely be 6/7 to 7/7 2014.
There are only a few topologies/variants that I have not touched on
that would go well with the slew EF3.
1. IC based .... either a high Vcc based LMxxxx based one, or a level shifted
symmetrical crown/QSC IPS with 15V supplies/TL071-5532 IC.
2. A simple ASKA/DX bootstrapped LIN amp without any CCS's.
Just simple resistors and a cascoded (low voltage)/ regulated LTP.
The regulated lower voltage would preserve PSRR without CCS's !
3. The "superpair" KODA type amp (below). Apex (audio) uses the
symmetrical version and Koda ... just a single ended version.
There are others ... but most fall into the "variant" category.
A fet IP pair "john curl" type IPS is the only other unique one.
Of course , with an error correcting or HEXFET OPS , there are
even more options
.PS- almost forgot .... a "wolverine" type input stage running a
Hawksford with a level shifter (like the original Luxman 120a -below 2) ...
might also make for listening pleasure.
OS
Attachments
Last edited:
I forgot to ask this and maybe it was written somewhere amongst the 2600 posts :O
How quiet are the CFA amps? the CXA and CXAH specifically?
I've got a sensitive pair of mid rangers that really hiss like mad if the noise level isn't down low. The wolverine is fine, as is no surprise, but how do those CFA amps compare? And what is the comparative distortion of the CXA and H at say 1kHz?
I am just trying to narrow down which IPSs give what specs. I know the H will probably give better numbers than the version without the more complex VAS, but you lose voltage headroom. As I'll be running these on ~+-25 volt rails, I need to be careful to pick the right amp. I know the wolverine will work fine, but I really want to try something 'different' having lived with a blameless on my mid/tweets for years.
And one final question. The wolverine sims far better at 10kHz than I am actually measuring. Is this simply a case of real life vs the simulator? I mean I'm getting noise floor levels of distortion prior to the feedback rolling off, but both 2nd and 3rd order hit around 0.002% by 10k. This to me would seem to be pretty great anyway, but if there's room for improvement I'm game to try tweaking stuff.
How quiet are the CFA amps? the CXA and CXAH specifically?
I've got a sensitive pair of mid rangers that really hiss like mad if the noise level isn't down low. The wolverine is fine, as is no surprise, but how do those CFA amps compare? And what is the comparative distortion of the CXA and H at say 1kHz?
I am just trying to narrow down which IPSs give what specs. I know the H will probably give better numbers than the version without the more complex VAS, but you lose voltage headroom. As I'll be running these on ~+-25 volt rails, I need to be careful to pick the right amp. I know the wolverine will work fine, but I really want to try something 'different' having lived with a blameless on my mid/tweets for years.
And one final question. The wolverine sims far better at 10kHz than I am actually measuring. Is this simply a case of real life vs the simulator? I mean I'm getting noise floor levels of distortion prior to the feedback rolling off, but both 2nd and 3rd order hit around 0.002% by 10k. This to me would seem to be pretty great anyway, but if there's room for improvement I'm game to try tweaking stuff.
CFA-X (h) are both single stage CFA's that use just the single pair for
the high Z input and the low Z feedback.
NAD "splits" the job in two ... first stage is a high Z diamond buffer running
the low Z current feedback 2'nd stage.
PS- CFA-X is faster .... 2 stages increase group delay -(lower slew).
CFA-XH is fastest still .... hawksford feedforward (no early effect)
seems to increase speed even more (compared to "X").
OS