Discrete Opamp Open Design

What do you want.... glory, fame or fortune? Pick one. Which one requires an advanced degree to obtain it?

Many a people with glory, fame & fortune have od'ed or committed suicide. I'll settle with happiness.

For some people the pursuit of happiness (nice movie, btw) involves fortune (or so they think), for others finding out "what makes things tick", for others getting a decent job, for others having a good quality of life, for others leaving their mark/making the world a better place/whatever etc.. Then again, others may be perfectly happy working as bouncers at clubs.

Of course nothing "requires" a degree. We require air, food, water and shelter (well, and the internet :p). It's just that you have better chances achieving some of the above with an advanced degree.
 
For some people the pursuit of happiness (nice movie, btw) involves fortune (or so they think), for others finding out "what makes things tick", for others getting a decent job, for others having a good quality of life, for others leaving their mark/making the world a better place/whatever etc.. Then again, others may be perfectly happy working as bouncers at clubs.
Hmmmm ... Christopher Langan - Wikipedia, the free encyclopedia ...

Frank
 
working as bouncers

By which one gets to hook up with the glorious, famous, and/or fortunate.

From successful writers to top crime attorneys, one-hour talks with hookers or accomplished actors/actrices.
From Diego Maradonna in deep coke, to smoking goldplated-filter cigarettes with a prince of Quatar.
From small talk about the weather with the CEO of Venezuela's national oil company, or billion dollar "crude" oil Texans with Bentley/day creditcards.
Or attempting to convince the No1 president of Daimler Benz that it's not opportune to order half a dozen callgirls
(by pouring him and his DB mates a glass of whisky).

Best ever chat I had lasted a couple of hours, with the public relations CEO of Nike, amazing individual.
(I've got money in the bank, of an advanced shoe sole patent, from Nike. My g/f's doing, not mine)

Pays well too.
Plus grand tips, plus gratuity for every taxi ride, nightclub or casino transfer, hooker stop-over, joint or coke ride.
Plus pays university tuition (at expense of a couple of added scars and tissue).
Two advanced degrees for the ticket of one.

(afraid I'm just a private boys boarding school bloke, served hard time with the priestly brothers at a catholico jail, they invited the antichrist in. Dennenstraat 10 in Siltech city, same joint as this guy : http://nl.wikipedia.org/wiki/Charl_Schwietert )
 
Last edited:
By which one gets to hook up with the glorious, famous, and/or fortunate.

One of my best pals is a bouncer.
As you probably know there are some less... interesting and/or fun parts on the job description (there is an abundance of red-necks with guns here), but it's generally OK and it pays well.
By the way, having (relevantly) advanced education and being fairly well brought up (his upper-class family consists mainly of doctors and high school teachers), he easily stands out from his peers (i.e. the average brute) and, more or less, manages two clubs now.

We're way off topic, though. Let's not turn this interesting topic into another Blowtorch thread! :p
 
Folks, just a brief interruption from the regular scheduled program: Here's a 10-BJT single-ended variant of a JLH gain-block design posted earlier on this thread or elsewhere on this forum.

The JLH was FET-input, symmetrical and had a cascoded VAS. I have added the PNP buffer (aka Blameless/JE990), and retained the LTP current-mirror. The new trick is a matched differential current source for the LTP and the VAS, with a 1:1 current ratio between the LTP (combined legs) and the VAS.

Various other circuit elaborations are possible - cascoded VAS, cascoded JFET LTP, etc., but the minimal circuit shown above simulates pretty well as-is. It appears to work OK at fairly low rails as well as low bias currents in the gain block - down to 200 uA and below.

Edit: Added missing attachments. THD20 FFT shown is for 4V output amplitude into 600R - about -105 dBr, mostly H2 with the Class-A biased output stage shown.
 

Attachments

  • jlh_se_bjtop.jpg
    jlh_se_bjtop.jpg
    81.4 KB · Views: 435
  • jlh_se_bjtop_4v600r_fft.jpg
    jlh_se_bjtop_4v600r_fft.jpg
    63.7 KB · Views: 435
Last edited:
Do you have a circuit that will be -140 THD N ? Or greater than -130?

A modded FET990 can get to the region of -130 dBr with some or all of the following mods:

1) Separate LTP and VAS CCSes.
2) JFET Cascoded LTP CCS.
3) Wilson CM for the LTP.
4) Cherry compensation.
5) Class-A biased output stage.
6) Ultra-low noise JFETs for the LTP - InterFET 9030 or similar.

I have DIP8 FET990 that implements 1, 2, 4 and optionally 5 in the works. It's all-SMD though, so that limits the choice of LTP JFETs to the BF862, 2sk209 and a few Sanyo/ON Semi low-noise devices.

Edit: I forgot the most important mod:
7) Omit the VAS anti-latching diode or use an ultra-low leakage diode like a BAV199.
 
Last edited:
Disabled Account
Joined 2012
A modded FET990 can get to the region of -130 dBr with some or all of the following mods:

1) Separate LTP and VAS CCSes.
2) JFET Cascoded LTP CCS.
3) Wilson CM for the LTP.
4) Cherry compensation.
5) Class-A biased output stage.
6) Ultra-low noise JFETs for the LTP - InterFET 9030 or similar.

I have DIP8 FET990 that implements 1, 2, 4 and optionally 5 in the works. It's all-SMD though, so that limits the choice of LTP JFETs to the BF862, 2sk209 and a few Sanyo/ON Semi low-noise devices.

Edit: I forgot the most important mod:
7) Omit the VAS anti-latching diode or use an ultra-low leakage diode like a BAV199.

If avail to others, I'll try it and also what about the SW-OPA design.... if they ever stop fooling around with smd and pcb and the like... and finally get one built and tested.... need to see actual results as opposed to SIM only. Looking to push things 5-10dB better.

-RNM
 
Last edited:
Do you have a circuit that will be -140 THD+N ? Or greater than -130?
The circuit I posted at #2060 should do this. The attached curve is distortion @ 20kHz 10Vp into 600R. About -104dB for 2nd.

The gain is 40dB. This is in fact similar to the "distortion magnification" method that the uber makers use to get numbers for their uber OPAs. So I expect 2nd to be less than -140dB at unity gain. eg fig2 LME49710 datasheet

Of course, at these levels, decoupling, layout, earthing & other factors have as much, if not more, influence on THD. eg C5, which ostensibly is to reduce +ve PSR will also reduce THD in real life .. even with supa dupa power rails.

I pontificate further at
A modded FET990 can get to the region of -130 dBr with some or all of the following mods:

1) Separate LTP and VAS CCSes.
2) JFET Cascoded LTP CCS.
3) Wilson CM for the LTP.
4) Cherry compensation.
5) Class-A biased output stage.
6) Ultra-low noise JFETs for the LTP - InterFET 9030 or similar.

7) Omit the VAS anti-latching diode or use an ultra-low leakage diode like a BAV199.
I don't think you will gain anything with 1) 2) 3) or 6)

My Q9 makes it easy to select output bias current. The sim is for 7mA so Class B for 10Vp 600R

But you can't do 7) Omitting your D6 gives really nasty overload. BAV199 won't do much to help. It's not leakage but modulation of the junction capacitance that introduces the distortion. Unfortunately, ANY diode in that position will give THD only about that of the original 990. :(
 

Attachments

  • 2060distortion600R.gif
    2060distortion600R.gif
    33.4 KB · Views: 373
Last edited: