DCAudioDIY.com

DC Area Audio DIYer's Community
It is currently March 28th, 2024, 6:46 am

All times are UTC - 5 hours [ DST ]




Post new topic Reply to topic  [ 18 posts ]  Go to page Previous  1, 2
Author Message
PostPosted: December 19th, 2013, 5:48 pm 
Offline

Joined: March 5th, 2013, 9:35 am
Posts: 259
Location: Highland, MD
The AD825 op amp I used is specified for rails of +/-5 to +/-15 Vdc, so a 24-V rail is not recommended (unless you like to see the Magic Smoke.)

_________________
- Guy


Top
 Profile  
 
PostPosted: December 19th, 2013, 6:03 pm 
Offline
User avatar

Joined: December 14th, 2013, 2:19 pm
Posts: 948
Thanks Tom.

The circuit used actually grounds the + rail of the op amp, so it's only "seeing" +/-6V now, +/- 12V with the higher potential supply.

Stuart


Top
 Profile  
 
PostPosted: December 19th, 2013, 6:06 pm 
Offline
User avatar

Joined: December 14th, 2013, 2:19 pm
Posts: 948
Remember, the object was to bring the DC output of the CF to 0. So, the op amp, MOSFET and references are all tied to the (-)V supply, that's where they have "room" to adjust.


Top
 Profile  
 
PostPosted: December 19th, 2013, 6:15 pm 
Offline
User avatar

Joined: December 14th, 2013, 2:19 pm
Posts: 948
Here is the circuit as originally drawn by Broskie. Guy and I made a few changes based on experience with implementation. NO connection to B+, you can actually get over 32V on the output during warmup!!! We connected to a low voltage + supply (6V3) and never had more than a couple tenths of a volt during warm up or turn off. 0.1uF caps didn't get the job done. Poor bass response. Switched the FB cap to 3.3uF (cause I had them) and the bypass cap to 0.47uF and all was hunky dory. Switched out the 7912 for an LM337 per Yaniger's design and the AD825 Guy designed for. Filaments are not connected to this supply in my boxes.


Attachments:
CIRCUIT3.gif
CIRCUIT3.gif [ 11.11 KiB | Viewed 18878 times ]
Top
 Profile  
 
PostPosted: December 20th, 2013, 7:58 pm 
Offline

Joined: March 2nd, 2013, 2:43 pm
Posts: 216
Location: Potomac, MD
Stuart,

Couple of comments. Change the 1N4001 to something like a 1N4148. No reason to have added capacitance of a power rectifier when a small junction will do. I don't know what MOSFET you are using, but I would choose a bipolar transistor such as a MPSA42 as it has very low capacitance in this application. This transistor is a small TO92 plastic device with 300 volts capability.

I would use a follower with a wide dynamic range to drive each filter. Each follower should have a dc servo to eliminate the coupling caps. By wide dynamic range, you should consider a negative voltage supply of -60 volts and use the current sink transistor like you show in your circuit. The easiest way to make this work, though, is to go back to wrapping the feedback back to the grid of the tube. Otherwise, you have to get into precision voltage dividers and referencing the Op amp to the - 60-volt supply. You do need to watch the power dissipation on the transistor if you are pushing 7 mA as this would be almost half a watt in the transistor.
The reason for the follower driving the filter is that when you come off the plate into the filter, you end up with a non-linear source impedance for your filter--not a good thing. I use this filter buffering in my better preamps now for EQ, and it is a big improvement in filter accuracy.

David


Top
 Profile  
 
PostPosted: December 20th, 2013, 8:06 pm 
Offline

Joined: March 2nd, 2013, 2:43 pm
Posts: 216
Location: Potomac, MD
Stuart,

You could reference the Op Amp to ground if you use some sort of level shift transistor to drive the current-sink transistor. Then you you could go with the general topology you have. This would be preferable. The only thing is to maintain stability as you are adding an additional gain stage in your already fragile feedback loop. But I know it can be done, as I did something like this in my TF-12 preamp years ago.

David


Top
 Profile  
 
PostPosted: December 22nd, 2013, 10:47 am 
Offline

Joined: March 2nd, 2013, 2:43 pm
Posts: 216
Location: Potomac, MD
One other comment. You would not need the extra buffering if the impedance of the filter is much higher than the effective impedance at the plate of the amplifier tube. If this impedance was 10k and the input resistor to your filter was 500k, you would not need it.

The effect of not having the buffering and needing it is most easily demonstrated by building a circuit without the buffering and then applying a square wave through an inverse network filter and into your circuit. You should be able to recover the square wave without distortion as you apply progressively larger signal levels. You will need the buffering if you see the square wave becoming tilted in a non-symmetric way. It kinda looks like you need one set of filter components for the positive half of the square wave and another set for the negative half. I use an inverse RIAA network when I am making a phono circuit, and I need the extra buffer to make things come out right at high signal levels. This applies to passive EQ. Feedback EQ has its own set of issues, the biggest one is that the bass accuracy is dependent on the gain of the tubes because there is never enough gain to apply enough feedback to maintain accuracy in the bass. And if you had that much gain and that much feedback, you probably would not like the way it sounded anyway. So with tubes, the best overall accuracy, I feel, is had with passive EQ and buffered EQ networks. And this is tube independent. Again, you want high dynamic range on the driving to your filter. Otherwise you will have poor overload characteristics. Feedback circuits have the advantage here.

David


Top
 Profile  
 
PostPosted: December 23rd, 2013, 9:36 am 
Offline
User avatar

Joined: December 14th, 2013, 2:19 pm
Posts: 948
David,


I'm not using the supply as drawn by John Broskie. I'm using an LM317/337 variant as laid out by Stuart Yaniger. No diodes.

Mosfets used have been IRFF220 and IRF820 or 840, Have to look up the #.

I've seen the filter inaccuracies from simulations. It was apparent that a low source impedance was needed to get the expected filter results. In fact, using Broksie's recommended filter scheme, which lacks any stage prior to the filter, adversely affects output of the preceding stage (in another piece of equipment, DAC, Phono stage,etc.) into the outputs of another filter, i.e., high pass & low pass. So, when I implemented my first filter, I did add a ~medium impedance gain stage (to be able to adjust for different amplifier gain levels) and used a high impedance filter. My rule was 1:10 minimum.

Since then, I found that a medium impedance filter sounds better (cleaner), i.e, 10k vs 100k, so now I need a 1/10 impedance gain stage.

But I'm trying to eliminate the line stage>filter by making it all-in-one, eliminating a series gain stage setup. I think a gain of 10 will suit any foreseeable amplifier combinations, so I was asking about an excellent sounding tube, with a mu of around 20, that I can use as simple grounded cathode, to drive a filter network with a ~10k impedance. Obviously. there will be variance, since the filters will be:
-18dB/octave LP
-6dB/octave HP & -6dB/octave LP
-6dB octave HP

I understand that a follower here will have much lower impedance that driving off of a plate, but I'd really like to avoid adding six more followers and servos.

Stuart


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 18 posts ]  Go to page Previous  1, 2

All times are UTC - 5 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 35 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group