switched attenuator for measurement

[PRR]

H-P tooled-up special switches.

You can use a lot of generic DPDT switches. Build a 1dB pad, and bypass. Continue 2dB, 3dB, 6dB, 12dB. While you can go further, a high-dB switched pad's parasitics add error. Build a good 20dB pad, 3 or 4 times. So like 9 DPDT levers in a row, and 27 lo-% resistors (now stupidly cheap in low power ratings).

Amateur radio geeks use this scheme, though more at 50 Ohms. Steal their plan and up all the resistors 12X.
http://lea.hamradio.si/~s53mv/dds/attenuator.html
https://qrpguys.com/qrpguys-31-db-step-attenuator
https://qrpguys.com/wp-content/uploads/2018/04/atten_assy_041618.pdf
http://www.arrl.org/files/file/Technology/tis/info/pdf/9506033.pdf

The strip-line and micro-resistor techniques are needed for VHF radio, NOT audio. Wire it compact and neat. Wire like a high-gain stomp-pedal: you won't get 120dB at 20KHz if output wires cross-back over input wires.

 

[David Kulka]

Agree with the HP 350 recommendation. There's one on ebay for $28 right now, that's a bargain. I think it's nominal 600 ohms but if your source is 50 ohms you can add a 550 ohm input resistor and optionally a 600 output resistor, so it tracks accurately. This is a precise high quality unit that will work better than anything you could build, without getting into a huge, tricky DIY project.

https://www.ebay.com/itm/Agilent-HP-350D-Attenuator-Set-5W-55V-600-O-DC-IMC/142249293882?hash=item211eb7f83a:g:GQQAAOSwo4pYfpA5

 

[Tubetec]

Thanks for that info Prr,
I'm starting to understand the various types of attenuator and their effects on input and ouptut impedences.
I can see how high values of attenuation demand resistors which cause stray capacities to have increased effect.
I got a bag of 3200 resistors ,metal film 1%(ish) 20 of each value from a fraction of an ohm to 20Mohm, just a few dollars including postage from China.

I did one time make a resistor string att on a 12 way Lorlin ,It all added up to 1Mohm,I started off with around 250k resistor and gradually reduced the resistance of the sucessive ones, I didn't do any math whatsoever ,I simply adjusted each resistor along the chain up or down as needed to get what seemed to be an even step in volume with each click and a final mute position , absolutely 'on the button' db's of attenuation doesnt really matter so much for recording ,the simple repeatabillity of steps comes in handy though. It works great on a pickup output into a di  to fine tune your levels, as I often found the usual 20/40db attenuation switches on di boxes  are too much when say a hot pickup output is just clipping the input of the DI, you only need to shave off a handfull of db to be hitting it right on the money,a -20db pad dumps s/n , setting the guitar volume control at some arbitrary  amount down from top means no repeatabillity in terms of level .I sometimes use the same attenuator in reverse as a load matcher after an input transformer ,1meg being the internally set value of input impedance of most of my stuff. Tuning the load on the transformer secondary allows me to set the amount 'magnetic core ' I get in the sound ,if you have several transformer coupled stages with carefully adjusted relative gains/impedences ,so only the peaks are just causing barely noticable distortion ,you have a handfull of db headroom in the digital domain ,so that by the time your hitting 0db digital your program material has already been limited by the action of the previous circuitry and transformers to a high degree. Any sibilant or plosive sounds your vocalist issues at full tilt hit the transformers a wallop and get eaten up ,where an op signal path to digital is going to need miles more headroom to reproduce the odd transient that comes along ,course you patch all that up in the mix by adding plugin dynamics later ,but your signal has already suffered bit munch at that stage anyway..

Im getting a better feel for the attenuators now thanks for all of your contributions once again , I can see something around 600 on the input ,with a chain of lower values in  the ten of ohms range might suit my purposes.
I see some guys who have designed say 24 way attenuators , I see a range between  -10 and -20 I like the look of,can  I simply pluck that section of the attenuator out and build it on my own switch ? adding the combined value of the rest of the chain at the end to ground ?

Hi David,
Yes I noticed the abillity to calibrate impedence of the HP with external values of resistor as well , 28 dollars is a total steal ,even if it was to sit on the shelf as an ornament and only be used once in a while.  Be a handy item in studio just to tweak your gain structure to get those last few Db signal/noise on special sources  like vocals etc.

I guess this tread has some crossover with next doors DI post , loads of great info on impedence matching ,loading(or lack of) etc ,easily one of the most fundamental things to getting a good sound and the one of the most misunderstood. as well.
For me the LCR meter has made a huge difference ,for instance being able to look at your grid wirings stray capacitance directly ,normally your head would be boggled my math to figure it out approximately ,where you just hook up the LCR and bam you have it ,you can easily check how much neater shorter wiring impacts stray capacitance .

 

[chops]

Agree with the HP 350 recommendation. There's one on ebay for $28 right now, that's a bargain. I think it's nominal 600 ohms but if your source is 50 ohms you can add a 550 ohm input resistor and optionally a 600 output resistor, so it tracks accurately. This is a precise high quality unit that will work better than anything you could build, without getting into a huge, tricky DIY project.

https://www.ebay.com/itm/Agilent-HP-350D-Attenuator-Set-5W-55V-600-O-DC-IMC/142249293882?hash=item211eb7f83a:g:GQQAAOSwo4pYfpA5

Would an attenuator like this be useful for attenuating microphone output levels for feeding inputs of fixed gain mic preamps like V72's etc?  As in, would it present a consistent impedance load to the the microphone?

 

[Tubetec]

Looks to me like a perfect studio companion , I also run very simple preamps, some with no onboard gain control , a unit like that would suit me very nicely , 600ohms loading probably wont suit every situation ,but you can insert known resistances to modify the impedences , looks like it would be easy to add Xlr's to that unit for convience also.

 

[JohnRoberts]

Would an attenuator like this be useful for attenuating microphone output levels for feeding inputs of fixed gain mic preamps like V72's etc?  As in, would it present a consistent impedance load to the the microphone?

That would not deliver best S/N... The fixed gain mic preamp would add constant noise N while you attenuate the S, making S/N ratio worse.

Better approach is variable gain in mic preamp (that reduces preamp N with the S)  for optimal S/N.

JR

 

[Tubetec]

Morning John,
But surely if you attenuated the preamp outputs s/n ratio would be preserved ,attenuating the input you would sacrifice db.

 

[JohnRoberts]

Morning John,
But surely if you attenuated the preamp outputs s/n ratio would be preserved ,attenuating the input you would sacrifice db.

Sure, assuming you have unlimited headroom in the preamp. Gain trim us useful to keep signal above the noise floor and below clipping.

JR

 

[chops]

Just looking for a solution for using condenser mics on louder sources with fixed input gain mic pres (V72's)  to avoid clipping but preserve impedance relationship of mic to preamp.  (Currently using inline pads, which drop level at a fixed 15db (too much for some applications) and effect impedance.  (Not looking for output trim.)

 

[Tubetec]

Unlimited headroom , not really possible I suppose,
but with a nice simple triode based input section with zero feedback its very easily possible to drive it into % points thd at the output and its still a useable signal ,alltogether different than an op amp where things fall apart quick.

Ah right I see Chops,
why not make your own 5 and 10 db inline pads .Is is a shame to have to scrub off 20db ,when maybe 5 db is enough to do the job.