Trasformerless (class A) DOA mic preamp - design discussion

Help Support GroupDIY:

Dualflip

Well-known member
Joined
Nov 15, 2008
Messages
1,401
Location
México City
abbey road d enfer said:
Do you know the value of rbb' ? I don't see any noise specs for this.

The primary purpose of those transistors is not low noise amplifiers so the data sheet doesn't show anything about noise, they just happen to be really quiet.

The value is around 1.72 ohms, it was measured and presented at:

http://www.dicks-website.eu/low_noise_amp_part3/part3.html

 

moamps

Well-known member
Joined
Sep 24, 2005
Messages
1,811
Location
Croatia, HR
Aleguitarpro said:
About cables I use generally Mogami, Sommer or Reference ("ultimo cavo" is my latest addition and in italian means literally "latest cable"... )...
I haven't mentioned GML between my favourite because I haven't a direct reference on guitar.
I tried it with a violin and I found GML highs are a little too "piercing", if I can use this term.
But I haven't any direct experience on guitar, maybe it helps details and sounds great.
Thanks. Have you maybe considered any possibility of changing the input impedance of the preamplifier?

 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,955
Location
Hickory, MS
moamps said:
Thanks. Have you maybe considered any possibility of changing the input impedance of the preamplifier?
The nominal design target for mic preamp input Z is bridging or 10x the source impedance of the mic so typically 1.5k- 2k.

Of course some mics can sound different (not better) if mis-terminated. Mics with active electronics could see a level difference, dynamic mics with no active electronics could suffer frequency response variation.

JR 

PS: In the way back machine mics and pre amps were designed to work together. Modern designers design them independently so target nominal terminations.
 

moamps

Well-known member
Joined
Sep 24, 2005
Messages
1,811
Location
Croatia, HR
JohnRoberts said:
The nominal design target for mic preamp input Z is bridging or 10x the source impedance of the mic so typically 1.5k- 2k.
Of course some mics can sound different (not better) if mis-terminated.
Just for you, I took out the input impedances of some mic preamps that the OP mentioned here to test. What is the conclusion?

GML 8304 - 1k ohms
Millenia HV 35 - 2,210 ohms
VP312 -  1k2 or 3k ohms
NEVE 1074 - 300ohm or 1200ohm
Avalon 737 - 850/2500 ohm
Buzz audio Ma2.2 - 3k ohms/1k2
Forrsel SMP2 - 13k ohms
CHANDLER LIMITED Germanium Pre - 200 ohms
 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,955
Location
Hickory, MS
moamps said:
Just for you, I took out the input impedances of some mic preamps that the OP mentioned here to test. What is the conclusion?

GML 8304 - 1k ohms
Millenia HV 35 - 2,210 ohms
VP312 -  1k2 or 3k ohms
NEVE 1074 - 300ohm or 1200ohm
Avalon 737 - 850/2500 ohm
Buzz audio Ma2.2 - 3k ohms/1k2
Forrsel SMP2 - 13k ohms
CHANDLER LIMITED Germanium Pre - 200 ohms
I have even written about this here before too... When marketing exotic, esoteric SKUs that you claim are superior to standard SOTA  products, it always helps to "sound" different.  ::) Extreme deviation from nominal "bridging" input impedance terminations can deliver a somewhat different sound, that a slick marketer will capitalize on... "see it sounds different".

If lower than nominal input impedance was somehow useful (better) the mic manufacturer could easily include a resistive termination inside the mic for only pennies. Unfortunately that would also give them a worse sensitivity spec so undesirable for specification readers.  Higher than 10x termination is not expected to make a huge difference (by me) but the mic manufacturer could ask for that if it provided a significant benefit (perhaps a slightly hotter output). Modern solid state preamps have an input impedance determined by passive components so are pretty much arbitrary.

I get your point that perhaps the OP could experiment with input terminations to get different sounds ... sorry if I stepped all over your suggestion.

JR

PS: Back last century I chased down a senior microphone design engineer at Peavey (I think he worked for Shure before working at Peavey). I asked him what kind of mic preamp termination would be ideal or best for him... First he looked at me like I was crazy, then answered that they design for nominal terminations and don't experiment with others because what is the point, for typical applications. 
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,976
Location
Marcelland
JohnRoberts said:
answered that they design for nominal terminations and don't experiment with others because what is the point, for typical applications.
It's a little too bad. I think he missed something. Some designers have capitalized on designing high-ish input impedance in mic preamps. Indeed the difference is subtle at best between typical 1.5-2k and higher (5-10k). But there is a significant difference between bridging and matching. In particular, the 300R setting of Neve preamps is a real tone-killer on some dynamic mics.
Now Peavey may not have been the best playground for such endeavours...
 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,955
Location
Hickory, MS
abbey road d enfer said:
It's a little too bad. I think he missed something. Some designers have capitalized on designing high-ish input impedance in mic preamps. Indeed the difference is subtle at best between typical 1.5-2k and higher (5-10k). But there is a significant difference between bridging and matching. In particular, the 300R setting of Neve preamps is a real tone-killer on some dynamic mics.
Now Peavey may not have been the best playground for such endeavours...
I was hired to work for AMR Peavey's recording (cough) division... The guy who recruited me  wanted to model classic recording mics with lower cost "soundalikes" (you can imagine the obvious list).  It was actually a merchantable idea but the technology back in the mid 80s, and my only analog design chops were not up to that task. The even larger problem was that AMR sales distribution were mostly Peavey dealers too, so not quite ready for "hip" recording products.

Within a year my boss, who recruited me to move to nowhere MS, was fired and I ended up transferred into Peavey headquarters to run mixer engineering for all of Peavey.

It's not all that hard to make a circuit sound different... making it sound different and better is much harder (IMO).  8)

JR 
 

Dualflip

Well-known member
Joined
Nov 15, 2008
Messages
1,401
Location
México City
I remember prefering the 500 ohm setting on the UA LA-610 rather than the 2K setting, but I'm probably attributing it to a change in level, the 500 ohm sounded louder which doesn't make sense since a 500ohm input will have less voltage than a 2K input as it acts as a voltage divider. So something else must have been going on there rather than just an resistor switch.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,976
Location
Marcelland
Dualflip said:
I remember prefering the 500 ohm setting on the UA LA-610 rather than the 2K setting, but I'm probably attributing it to a change in level, the 500 ohm sounded louder which doesn't make sense since a 500ohm input will have less voltage than a 2K input as it acts as a voltage divider. So something else must have been going on there rather than just an resistor switch.
With a typical 150-200R mic, the divider effect is still in favour of the lowest impedance setting.
Let's say mic Z is 200R, input Z is 500R/2k
With the 2k position, A=0.9 or -1dB
With the 500R position, A=0.71 or -3dB, but there is 6dB more due to the turn ratio.
So, clearly, the 500R position wins in terms of sheer level, with an advantage of 4dB.
This difference is largely enough to fool the brain into thinking "better". This is normal when we think of audition as a means of detecting unseen enemies or dangers. Typically, I've found that a mere 0.5dB difference in level is perceived as "better" by non-trained listeners.
Only when levels are normalized can we really assess the more subtle differences.
I first noticed the unfavourable effects of too much loading (too low input Z) when I bought a Groove Tubes "The Brick". I thought it was extremely dull, and I traced that to the 500r input impedance, due to the hybrid NFB (voltage and current) in the input stage. I was shocked when I learnt that Terry Manning (Ocean Ways) uses it quite often, albeit on non-essential tracks.
Of course, with most condenser mics, the effect of loading is minimal, but very noticeable on dynamics and ribbons.
 

Bo Deadly

Well-known member
Joined
Dec 22, 2015
Messages
2,610
Location
New Jersey, USA
abbey road d enfer said:
I first noticed the unfavourable effects of too much loading (too low input Z) when I bought a Groove Tubes "The Brick". I thought it was extremely dull, and I traced that to the 500r input impedance, due to the hybrid NFB (voltage and current) in the input stage. I was shocked when I learnt that Terry Manning (Ocean Ways) uses it quite often, albeit on non-essential tracks.
Of course, with most condenser mics, the effect of loading is minimal, but very noticeable on dynamics and ribbons.
I aways thought making the signal "dull" was sort of the point (in addition to the noise benifits). If by "dull" you mean it's just removing highs, then great. If you're going to remove highs or lows, why not do it by not amplifying it in the first place? That's why I love the idea of making the cap in the gain control network of a typical mic pre too small as a way to low cut (even it you do have to compensate for Fc changing with gain setting). One of my all-time favorite EQs is a Jensen P12Q speaker. It's a brick wall filter above ~4K.
 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,955
Location
Hickory, MS
squarewave said:
I aways thought making the signal "dull" was sort of the point (in addition to the noise benifits). If by "dull" you mean it's just removing highs, then great. If you're going to remove highs or lows, why not do it by not amplifying it in the first place? That's why I love the idea of making the cap in the gain control network of a typical mic pre too small as a way to low cut (even it you do have to compensate for Fc changing with gain setting).
That was a cheat made by one large MI manufacturer who simultaneously spent $millions advertising how "quiet" (cough) their preamp was. That preamp only met its frequency response specification at lower gains, at max gain the undersized gain leg capacitor rolled off the very LF (like 1/F) noise.  In unscientific side by side listening tests it did indeed sound quieter, because it was (not flat).
One of my all-time favorite EQs is a Jensen P12Q speaker. It's a brick wall filter above ~4K.
When doing severe band limiting psychoacoustics researchers have found that it sounds more natural to trim response from both ends of the bandpass.  So a brick wall 4k LPF does not want to see 20Hz low frequency skirt.

JR
 

Bo Deadly

Well-known member
Joined
Dec 22, 2015
Messages
2,610
Location
New Jersey, USA
JohnRoberts said:
That was a cheat made by one large MI manufacturer who simultaneously spent $millions advertising how "quiet" (cough) their preamp was. That preamp only met its frequency response specification at lower gains, at max gain the undersized gain leg capacitor rolled off the very LF (like 1/F) noise.  In unscientific side by side listening tests it did indeed sound quieter, because it was (not flat).
Ah, and I thought I was being clever. But please name the product if there's a schem. I'd really like to see the details.

JohnRoberts said:
When doing severe band limiting psychoacoustics researchers have found that it sounds more natural to trim response from both ends of the bandpass.  So a brick wall 4k LPF does not want to see 20Hz low frequency skirt.
How much so? Like 20dB / oct starting at 90Hz? That's what the P12Q does already.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,976
Location
Marcelland
squarewave said:
I aways thought making the signal "dull" was sort of the point (in addition to the noise benifits). If by "dull" you mean it's just removing highs, then great.
I don't follow you. Removing highs is not my idea of a good signal path, because most of the highs is musically significant, as opposed to lows under 40Hz, that are consistently crap.

One of my all-time favorite EQs is a Jensen P12Q speaker. It's a brick wall filter above ~4K.
For taming the highs of a Telecaster, yes, but not for vocals, piano or cymbals.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,976
Location
Marcelland
JohnRoberts said:
When doing severe band limiting psychoacoustics researchers have found that it sounds more natural to trim response from both ends of the bandpass.  So a brick wall 4k LPF does not want to see 20Hz low frequency skirt.
That was known to me as the 400 000 rule (20Hz-20kHz, 40Hz-10kHz...). I've seen variants.
The key word is "severe band limiting"; when applied to a full-spectrum program, it is true that some kind of symmetry must be observed, but when it comes to balancing a program, it is not necessary. There are many examples of records that have almost inexistent bass and sound good. Not so with records with no highs.
 

Winston OBoogie

Well-known member
Joined
Jun 3, 2004
Messages
2,639
Location
UK.
Dualflip said:
I remember prefering the 500 ohm setting on the UA LA-610 rather than the 2K setting, but I'm probably attributing it to a change in level, the 500 ohm sounded louder which doesn't make sense since a 500ohm input will have less voltage than a 2K input as it acts as a voltage divider. So something else must have been going on there rather than just an resistor switch.

That particular product front end is based on the earlier 2-610 and M-610,  which I designed ( the word 'designed' doing a lot of heavy lifting there) 
The difference in input Z is derived from a dual tap on the input transformer which would, in the 500 ohm setting, result in a change from a 1:7 ratio to 1:14 ratio hence, the difference in gain.
 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,955
Location
Hickory, MS
squarewave said:
Ah, and I thought I was being clever. But please name the product if there's a schem. I'd really like to see the details.
I generally don't like to bad mouth competing brands by name, but this was all several decades ago. The mixer company was Mackie and I expect they knew what they were doing. (Greg and his lead engineer Rick Chin knew their way around mixer design from earlier mixer company gigs working in the industry). 

Another dubious "trick" they used in a 4bus mixer design that they marketed as having "high headroom" while they did not detect clipping in the +10dB post fader gain stage..  It was possible to experience severe audible clipping without any visible O/L indication. I taught my sales reps about these design flaws at seminars.

Later one rep told me that he walked into a dealer of his that was feeding one of these Mackie mixers into a Peavey amp and speakers (at this time Mackie only offered a few mixers, no amps or speakers.) My rep asked the dealer why he was so obviously clipping the audio. The Dealer answered that the "High headroom" Mackie mixer was creaming the Peavey amp that couldn't handle the hot signal output.  ::) My rep walked over and trimmed the channel fader down a few dB and master up the same few dB  to clean up the audio signal.  8)

Its funny but millions of dollars of advertising can buy you a lot of good customer expectations (even dealers that should know better, but don't). One could argue this was an innocent mistake that worked in their favor... I didn't consider them beginners who happened to make several lucky mistakes.

How much so? Like 20dB / oct starting at 90Hz? That's what the P12Q does already.
As abbey said its proportional.

This is more of a mixing thing than product design issue. After I dumped iron back in the 70s full bandwidth to both extremes was not difficult to deliver inexpensively.

JR

PS: I don't remember which exact model mixer did the undersized gain leg cap trick, IIRC it was during the massive VLZ preamp marketing campaign so one of those VLZ models (I didn't dissect every single one of their mixers and they didn't share schematics with me, but from time to time I would run a new model across my test bench to kick the tires.)  The 4 bus mixer with the missing clip detectors was a "SR" something series (I think... it was decades ago and I try not to fixate on how bad they kicked my ass in the mixer market). 
 

Bo Deadly

Well-known member
Joined
Dec 22, 2015
Messages
2,610
Location
New Jersey, USA
JohnRoberts said:
PS: I don't remember which exact model mixer did the undersized gain leg cap trick, IIRC it was during the massive VLZ preamp marketing campaign so one of those VLZ models
Well the schems are all on the Mackie site (which is fantabulous BTW and one of the reasons why I just recently stolebought a basically mint 1402VLZ4) and indeed I can see that early on the caps were a little small. In particular, the Micro Series 1402 VLZ used 4.7R / 470u for a cutoff of 72Hz. But later models it seems they abandoned the idea and went whole-hog for 3.32R / 3300u. Personally I think they should have kept the 470u. LC would only kick in at the very end of the pot travel.
 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,955
Location
Hickory, MS
squarewave said:
Well the schems are all on the Mackie site (which is fantabulous BTW and one of the reasons why I just recently stolebought a basically mint 1402VLZ4) and indeed I can see that early on the caps were a little small. In particular, the Micro Series 1402 VLZ used 4.7R / 470u for a cutoff of 72Hz. But later models it seems they abandoned the idea and went whole-hog for 3.32R / 3300u. Personally I think they should have kept the 470u. LC would only kick in at the very end of the pot travel.
When you publish a frequency response specification it is ASSumed to be worst case, not best case. They could have footnoted their mic preamp's frequency response spec sheet with an @XdB gain qualification, but did not AFAIK. 

In side by side WFO listening test of the mic preamp noise floor, a HPF at 70 Hz will sound much cleaner (quieter) than a full bandwidth design (like mine were). Add millions of advertising dollars touting how much better your preamps are and the sheeple believe the big lie (aka expectation bias). 

JR

PS: Not to pile on but I think I found an issue with the insert circuit on that model (1402?)... IIRC the insert jack feeds a Baxandall tone control with no buffer, so tone control settings will change the effective input impedance and interact with higher Z insert sources. I discovered this by accident when I noticed a measurable difference between 50 ohm and 600 ohms source impedance from my test bench 8903.
 

Bo Deadly

Well-known member
Joined
Dec 22, 2015
Messages
2,610
Location
New Jersey, USA
JohnRoberts said:
When you publish a frequency response specification it is ASSumed to be worst case, not best case.
OMG JR. Yes, Mackie kicked you in the balls. Get over it. If it makes you feel better, I sincerely doubt it had anything to do with the circuitry. The problem is that the Peavey brand never got a ton of respect from the pro-audio community. People are fickle. They buy stuff because of how it looks. How it works is important. It has to work. But as long as it does, it's actually secondary to how it looks. If they just changed the logo alone that would have boosted sales literally 100%. The Peavey logo looked out-dated by ~1986 (and IMO it was never cool). I grew up in the 80's when Peavey stuff was high profile in the stores. I still have a Renown 400 buried in my mom's attic somewhere. My impression was that Peavey stuff was popular for PA applications and cheap guitar amps. If you needed some watts, you get some Peavey stuff and put it in the background. People also get bored easily. So some new thing comes out and all of the sudden it's all-the-rage. You never had a chance.
 

Latest posts

Top