Hardy M1 type build

[seavote]

I've taken the M1 schematic available at john hardys' web sight and stripped away the ground lift,phase switch etc and came up with the schematic posted below.

Schematic edited 10/23/
10/27

will the circuit work as drawn??
the original can be found here:
http://www.johnhardyco.com/Schematics/MPC-1_Schematic.jpg

this schematic is linked by me to the prodigy pro audio forum with Mr. Hardys' permission on a one time basis.it may not be altered in anyway or used on any other forum without Mr. Hardys' permission. permission is not given to copy it to any other location and link to the schematic there.

i've kept the component numbers in my stripped down schematic the same as in the original to avoid confusion when comparing the two. i'll be using an mnats 990 from one of ptowns kits, a stepped attenuator as in the hardy M2 and U2 will be an analog devices OP97. i just relized i did nothing with the OTs' (-) leg on the primary side.should this just go to ground? thanks for any help and /or suggestions

 

[AMZ-FX]

[quote author="seavote"] i just relized i did nothing with the OTs' (-) leg on the primary side.should this just go to ground? thanks for any help and /or suggestions[/quote]

Yes, it goes to ground through another of the Jensen resistor/inductor packages (RL2).

regards, Jack

 

[John Hardy]

The low side of the primary of the output transformer can go directly to ground, but the path through the resistor/inductor combination to ground provides a more balanced output impedance. This can result in better common mode rejection at the receiving end.

John Hardy
The John Hardy Co.

 

[seavote]

thanks. i edited schematic to reflect suggestions. i see on the john hardy co web site the dc offset calibration for this circuit requires a meter with a sensitivity of 1 micro volt. i dont have one or, access to one. should i remove the trimmers from the circuit or do the best i can with what i have. i can get my hands on a good fluke meter.
would i be doing more harm than good calibrating without the proper voltmeter? also i'm assumming the rotary switch for the attenuator is a shorting switch. does that sound right?

 

[John Hardy]

The circuit should work very well, assuming you lay out the p.c. board properly.

RV3 fine-tunes the DC-servo circuit op-amp to achieve the lowest possible DC offset of the 990. If you simply delete RV3, the OP97FP op-amp will provide very good results, probably much less than 200 uV for the DC offset of the 990. This is generally excellent, but the addition of RV3 and using a meter capable of accurately measuring DC voltages down to 1 uV or so can obviously improve it. Even then, you have to properly "zero" the meter so that you aren't adding the inherent errors and drift of the meter itself to the measurement. So, you might be better off just leaving the trimmer out. Design a place for it, but don't install it.

RV2 adjusts the input bias current compensation of the 990. This does not require a meter with 1 uV sensitivity. Probably 100 uV sensitivity is fine. You will be measuring voltages as low as 1 mV or so, so a meter with 100 uV sensitivity is sufficiently accurate. If you don't have one, you should not be doing this project.

John Hardy

 

[John Hardy]

The 16-position switch for the gain control of the M-2 is a shorting switch. If it were a non-shorting switch, the resistance from the inverting input to ground would be infinity for a brief period as the moving contact of the switch went from one position to the next and was not connected to either stationary contact (thus, non-shorting). This would cause the 990 to be operating at unity gain for that brief period. Although the 990 is considered to be unity gain stable, it is not recommended to operate it at unity gain. It is the least stable mode.

This switch is not an "attenuator". It is being used as a rheostat (a 2-terminal variable resistor) to adjust the gain of the 990.

John Hardy

 

[JohnRoberts]

It's good to hear from the Mr. Hardy when talking about the Hardy pre.

Congrats on a well respected design that holds up like a fine wine.

I wonder a little about how import the fine trims are, but there's nothing wrong with making things as good as possible.

JR

PS: For Seavote, the marking around the phantom switch looks odd. I gather that really isn't a 18- or 48- since it is also tied to ground.

 

[John Hardy]

[quote author="JohnRoberts"]I wonder a little about how import the fine trims are, but there's nothing wrong with making things as good as possible.[/quote]

Originally my MPC-600 and MPC-3000 cards did not include the servo trim pot. But I added it when I redesigned those cards several years ago. It is simply the addition of a 1k trim pot.

There are op-amps that are specifically designed for very low DC offset and drift, such as the LT1012, LM11, AD705 and OP97. There are all kinds of applications that require great DC specs. The trade-off is that the slew rate is usually very low. These four op-amps (and at least some others) are designed to allow for the addition of a trim pot connected with one end of the pot going to pin 1 of the op-amp and the other end of the pot going to pin 8, with the wiper of the pot going to the positive supply. This provides a simple and low cost way to trim the DC offset of the op-amp to provide the lowest possible DC error.

PS: For Seavote, the marking around the phantom switch looks odd. I gather that really isn't a 18- or 48- since it is also tied to ground.

The original schematic that I drew and Seavote used as his reference (see his link) says "+48V GROUND" for that connection with "+48V SUPPLY" being the label for the supply line above it. Seavote decided to label the ground connection there as "48V-". The "4" does start to look like a "1". Personal style.

John Hardy

 

[seavote]

John H,thanks for calling my lack of attention to detail, "personal style". very kind of you. i've changed the schematic to reduce confusion.
i'll leave rv3 uninstalled. rv2 will remain,1mv sensitivity i can do.
i'm not sure i understand the rheostat-attenuator distinction but after some research im thinking its not about the hardware and components used its a distinction of function within the circuit. an attenuator would be used after the 990 gain stage to adjust the signal level?
John R, yes that part of the schematic was not presented clearly. those were/are 48s but part of the drawing was cut off while scanning to a usable pc file. didnt notice till you brought it up.

 

[JohnRoberts]

Rheostat is just a variable resistance (two terminal), while a potentiometer or attenuator is a divider (three terminal).

JR

 

[John Hardy]

John H,thanks for calling my lack of attention to detail, "personal style".

This is one reason why I told you in a separate e-mail to draw your own schematic. It's a good way to learn the basics, such as:

1. Getting the labels right (and component values, etc.).
2. Organizing the layout of the schematic so that (for example) the input transformer's secondary doesn't have to double back to get to the non-inverting input of the 990. It doesn't matter in a schematic, but it will matter in a p.c. board layout. But the clearer it is drawn, the better it is for everyone.
3. Making certain that all the connections and parts are there, and you don't forget anything (like the low side connection of the primary of the output transformer).
4. You draw it clearly and consistently so that parts that are supposed to be connected are drawn as such, and parts that aren't supposed to be connected are drawn accordingly. Your use of large dots at some intersections, but not at others, is ambiguous.

This is good practice for the p.c. board that you will have to lay out. Same principles apply there, only they are more critical. Your hand-drawn schematic gets the job done (after some corrections), and there is the original schematic that people can refer to via the link that you provided in case there are any questions.

John Hardy

 

[seavote]

schematic has been improved but not perfected. i did learn some, as i always do with each new project.
i started some layout work on this a while back and have revisited it this week. i'm compartmentalizing the circuit and trying to do it a section at a time. difficult job. it seems like doing a word jumble - once you get a few components in the right place the others fall right into place. i just got eagle so i was thinking of doing the final layout with it.(i've just been doing it by hand) it doesnt strike me as very intuitive.but i'll stick with it till im capable with it.i have a lot of work ahead of me (fun, interesting work just the same) thanks for all the help. anyones suggestions or comments are welcome.

 

[valvehead]

[quote author="seavote"]i just got eagle so i was thinking of doing the final layout with it.(i've just been doing it by hand) it doesnt strike me as very intuitive.but i'll stike with it till im capable with it.[/quote]

I'll admit that the leaning curve for Eagle is pretty steep, but once I caught on I wondered how I got by without it. I even ended up making my own device library almost entirely from scratch. If I only had an extra $1000 to get the professional version...