Talk to me about DIs

[midwayfair]

I need to build (or possibly buy) some DIs.

I will have access to phantom power for a couple of them, but I might need to resort to passive transformer for a couple.

Anyone have any suggestions? Projects, common transformers to use, or if this is a better-to-buy situation? I was trying to figure out if a transformerless solution like just using a Schoeps microphone circuit (with like 1M input impedance instead of 1G) would be good for this.

 

[rock soderstrom]

My advice for an active DI Box:

https://groupdiy.com/threads/bo-hansen-di-layout.36569/

 

[merlin]

In most gear the Di input is just a simple JFET source follower

 

[Newmarket]

In most gear the Di input is just a simple JFET source follower

Yes. But that doesn't give you a balanced output to drive a cable to another piece of kit as ina 'DI box'. I'm guessing that a "DI Box" type function is what the OP needs ? But yes a JFET follower gives you a high input Z and low output Z.

 

[merlin]

Yes. But that doesn't give you a balanced output to drive a cable to another piece of kit as ina 'DI box'.

Yes once you have buffered with a JFET you can do whatever you like with it. Balance with opamps, transformer, DRV134, or whatever you prefer

 

[L´Andratté]

The passive way is using a high ratio mic input transformer (1:10 or higher) in reverse (10:1, you probably knew that ). This has obviously often a "sound", which can be bug or feature, depending on you and the transformer.
In a pinch I have used a NTE-10/3 in an aluminum box as a budget version, I didn't check the freq response but hadn't any complaints with a Rhodes.
Also Sennheiser Tm003 1:15 xfrms can be found not too expensive.
One "holy grail" of iron sound seems to be (based on hearsay) the UTC A-10,
which is expensive and replicated by AMI/tab-funkenwerk, which is *surprise*
also expensive
Thinking about it, I should try that one myself for my J-bass...mumble, mumble...

 

[JohnRoberts]

I need to build (or possibly buy) some DIs.

I will have access to phantom power for a couple of them, but I might need to resort to passive transformer for a couple.

Anyone have any suggestions? Projects, common transformers to use, or if this is a better-to-buy situation? I was trying to figure out if a transformerless solution like just using a Schoeps microphone circuit (with like 1M input impedance instead of 1G) would be good for this.

This topic has been well discussed here over the years so my first advice is perhaps do a search (here).

The category affectionately known as DIs ranges from simple passive transformers to full blown preamps. Historically DIs were all passive and used transformers to scale down a source's impedance to be a better match to mic preamps (nominally 1.5-2k ohm). Passive transformers scale impedance by the turns ratio squared. Active DIs can use phantom power to operate without batteries (I designed one at Peavey back in the 80s using a bifet op amp inside).

One consideration is what kind of signal source are you trying to interface through a DI? Keyboards or gear with active drive outputs can easily drive a passive transformer DI without problem (a 10:1x transformer scales down source impedance 1/100x while scrubbing off 20dB of level). Some instrument signal sources benefit from the even higher input impedance available from active DIs.

It sounds like you are considering more than one application, so maybe consider more than one approach.

JR

 

[ccaudle]

Anyone have any suggestions?

tldr; go buy some

Longer explanation:
When thinking about building vs. buying here are some considerations in my opinion to help make the choice. Others may have different criteria, and probably most people will have different order of priority, so these are explicitly not in priority order.

• Is this something I want to do for fun? I think for most, or at least a lot of us here, DIY is something of an enjoyable diversion. If this would just be a tedious chore for you, that would be a mark against.
• Is this something not commercially available? If you want something which is not produced commercially, either because it is no longer trendy, or some of the parts needed are no longer available affordably in commercial quantities, or it is considered too labor intensive to be worthwhile now, then it might be a good choice for DIY. Examples I can think of are compressors based on old photocells that are no longer commercially available, or hand wiring tube circuits, restoring vintage devices, etc. Obviously doesn't apply to DI devices, they are standard items on every stage and you can walk into any pro audio or musical instrument store and buy a box full.
• Is this something I can do much cheaper than commercial devices? This is a difficult bar to hurdle for high volume devices, so mostly applies to boutique devices which may have a very high labor cost associated (e.g. hand wired tube amps), as well as a premium associated with being low volume (because then the company overhead charges get spread across fewer units).

The last one, can you do this cheaper, is especially tough if what you are trying to build comes in a simple metal box that can be punched out cheaply if you are making enough to justify tooling for automated hole punching and metal bending. If you are building something that vendors do with machining or hand bending you can maybe get close, but DI boxes are produced in the thousands at Chinese factories that have tooling setup to punch holes and bend the case on an automated line.

For comparison, spend about 15 minutes browsing the big online retailers like Guitar Center, Sweetwater, Sam Ash, etc. to see how much the low cost passive and active DI boxes cost at retail, then spend about an hour browsing for blank metal cases, transformers, connectors, etc.
My guess is that if you look pretty hard you might be able to find all the parts for not much more than you can buy a complete box for, but then you still have to spend at least an hour per punching or drilling connector holes, wiring up transformers, buffer circuits, etc.

That calculation may change if you want a high end boutique DI, e.g. maybe you want to duplicate a Rupert Neve Designs DI box, those probably have a premium associated with that "Rupert Neve" label above and beyond the usual component and manufacturing costs, but if a Livewire or Behringer works for your needs, I don't see any way you could beat their retail prices.

 

[midwayfair]

This topic has been well discussed here over the years so my first advice is perhaps do a search (here).

The category affectionately known as DIs ranges from simple passive transformers to full blown preamps. Historically DIs were all passive and used transformers to scale down a source's impedance to be a better match to mic preamps (nominally 1.5-2k ohm). Passive transformers scale impedance by the turns ratio squared. Active DIs can use phantom power to operate without batteries (I designed one at Peavey back in the 80s using a bifet op amp inside).

One consideration is what kind of signal source are you trying to interface through a DI? Keyboards or gear with active drive outputs can easily drive a passive transformer DI without problem (a 10:1x transformer scales down source impedance 1/100x while scrubbing off 20dB of level). Some instrument signal sources benefit from the even higher input impedance available from active DIs.

It sounds like you are considering more than one application, so maybe consider more than one approach.

JR

Hi, JR, I do generally try to search first. Unfortunately this was a hard topic to search for because "di box" brings up a lot of unrelated results. I found some results that were specific to passive or tube DIs but I was trying to get some general advice rather than assume I knew what would be best here (which I appreciate getting). Also, I'm not sure if this is a consequence of the forum software and you might be unaware, but search results don't actually go back that far anymore I just checked again and the oldest hit was 2020! Like I said, though, there were a couple threads about passive projects.

These are the signals I need to D.I. (not all at once):
keyboard with dual TS outs (this, as you said, should be fine knocking the signal down with a passive, but really I just would want to be able to run a balanced signal 50 feet to the board)
an acoustic with an active pickup -- 1:1, active, or any passive is fine here but would prefer not to knock down the signal
a guitar, mandolin, and fiddle all with passive pickups (these all need active as I would expect they need ~10M)

tldr; go buy some

I appreciate this advice. I am certain I can't beat the price on a Live Wire or Behringer, but I did think making a couple of them might be slightly more fun. I'm prepared to buy a couple if I don't like what I end up with.

ETA: Forgot to mention! A ton of the boxes I was looking at aren't even 1M input impedance, I was concerned about the passive instruments. Commercial DIs with very high input impedance tend to be expensive. I know I could just buy/build a preamp instead, but I'm trying to minimize the crap to keep track of on stage.

-----

I can at least mention what I have lying around:
Schoeps mic boards, which I could easily modify to 1M input impedance and have something running on phantom power with a balanced out. Just literally don't know if this will work for an instrument signal.
Several Edcor 600:600, they're a bit heavy and probably overkill but I could stick a FET follower in front.
THAT (1646 maybe? not sure exactly where I put them ...) chips that do balanced out, but I'm not sure I can run them on Phantom.
Some 10:1 but they're nice and I don't particularly want to use them.
Plenty of guitar pedal parts and some guitar pedal boxes that any purely active DI circuit would easily fit in.

 

[JohnRoberts]

Hi, JR, I do generally try to search first. Unfortunately this was a hard topic to search for because "di box" brings up a lot of unrelated results. I found some results that were specific to passive or tube DIs but I was trying to get some general advice rather than assume I knew what would be best here (which I appreciate getting). Also, I'm not sure if this is a consequence of the forum software and you might be unaware, but search results don't actually go back that far anymore I just checked again and the oldest hit was 2020! Like I said, though, there were a couple threads about passive projects.

as I mentioned "DI" has become a catch-all term describing many different functions than basic DI.

These are the signals I need to D.I. (not all at once):
keyboard with dual TS outs (this, as you said, should be fine knocking the signal down with a passive, but really I just would want to be able to run a balanced signal 50 feet to the board)

If you are plugging into mic inputs at the board a modest step down transformer will be fine.

an acoustic with an active pickup -- 1:1, active, or any passive is fine here but would prefer not to knock down the signal
a guitar, mandolin, and fiddle all with passive pickups (these all need active as I would expect they need ~10M)

The cheap active DI I did for Peavey offered a buffered unity output, but T-S not balanced. The balanced output was padded appropriately.

I appreciate this advice. I am certain I can't beat the price on a Live Wire or Behringer, but I did think making a couple of them might be slightly more fun. I'm prepared to buy a couple if I don't like what I end up with.

A criticism of the too cheap to not suck passive DIs, is inadequate shielding for the transformers to reduce cost. This may or not be a problem depending on how you use them

ETA: Forgot to mention! A ton of the boxes I was looking at aren't even 1M input impedance, I was concerned about the passive instruments. Commercial DIs with very high input impedance tend to be expensive. I know I could just buy/build a preamp instead, but I'm trying to minimize the crap to keep track of on stage.

-

this has been well discussed (more recently than 2020)... any active DI can use arbitrarily high resistive input termination.

----

I can at least mention what I have lying around:
Schoeps mic boards, which I could easily modify to 1M input impedance and have something running on phantom power with a balanced out. Just literally don't know if this will work for an instrument signal.
Several Edcor 600:600, they're a bit heavy and probably overkill but I could stick a FET follower in front.
THAT (1646 maybe? not sure exactly where I put them ...) chips that do balanced out, but I'm not sure I can run them on Phantom.
Some 10:1 but they're nice and I don't particularly want to use them.
Plenty of guitar pedal parts and some guitar pedal boxes that any purely active DI circuit would easily fit in.

Start making your own, you will learn from doing.

JR

 

[ccaudle]

I did think making a couple of them might be slightly more fun.

OK, that checks the first item on my list.

what I have lying around

If you are willing to consider your existing parts as sunk costs, or in other words not think of them as a new expense against this project, then it definitely helps overcome my third objection.

Schoeps mic boards, which I could easily modify to 1M input impedance

I don't see why that would not work. I'm no sure what the distortion profile is like at higher levels, so you may need to check how something like a really hard guitar strum or a handful of keyboard keys (i.e. highest level an instrument will make) sounds like through it.
And you don't have to stick to 1M, you could use 5M or 10M, see what works best for your instruments.

With a straight buffer type circuit (no transformer) you do not have an option for ground lift. I don't think that should be an issue for guitars, but may or may not be a problem with a keyboard. Just depends on how the keyboard power is wired up.

should be fine knocking the signal down with a passive

What is the gain structure on the board you are going to? If the minimum gain is low enough you may not need to cut the level.

Do you know the output impedance on the keyboards you have? At least a buffer is probably warranted unless the keyboard has a really nice output stage with low impedance already. I would at least try the Schoeps circuit, it might just work fine for your keyboards.

THAT...chips that do balanced out, but I'm not sure I can run them on Phantom.

Gets tough on phantom. Those chips really should be driven from a low impedance source, which means either a discrete or op-amp buffer in front of the output chip. The output chip pulls 5mA by itself, so by the time you add a buffer and account for current delivered to the load, it starts getting really tight to run on phantom power.

Several Edcor 600:600, they're a bit heavy and probably overkill but I could stick a FET follower in front.

Should work, but again if talking about a phantom powered buffer possible issues of not being able to separate input ground and output ground if the phantom powered buffer is in front of the transformer.
That may not matter for you, the band I run sound for uses a few DI boxes, and the only one I need to lift ground on is a DI fed from the effects send of a bass amp, so that I can take the output of the bass preamp direct rather than try to mic the amp. I've never had to worry about ground lift with the Roland or Yamaha keyboards we have used, or the acoustic guitar pickups.
That is kind of a philosophical decision, do you want to make something which is as flexible as possible, or is just meeting your immediate needs good enough?

 

[Kryss]

Hi! I'd like to butt in here too if that's okay, I'm pretty new to the group but I'm up for a similar project.

I'm building a studio and had the awesome idea to integrate DI in the XLR boxes with a switch to choose between active and passive. I found an older but (as it claims high quality and performance) schematic somewhere from a site that doesn't exist anymore where an active DI could be build using Phantom power/18VDC, also switchable. Could be that the Hansen is derived from that design, once I find it I'll post.

I do tend to build a 18VDC powersource for this then, the idea to use phantom from my pre-amps or console for this is a risk I'm not to keen on taking. A passive DI is there as well though, what seems like nothing much more than a couple resistors, capasitors and a 600ohm 1:1 (if I remember right) tranny to seperate the + and ground to hot/cold and ground. I'll go look for those schematics and upload them if you like if and when found.

Either additioned to this socket or another insert will be made for high Z equipment, I believe I have a schematic for these as well somewhere, with switches for the different impedances for different sources. Older guitaramps tend to get all over the place with those Z's.

Otherwise I'll find use for those butchered Behringer DI4000's where I needed the T1's from. Without those these machines aren't that bad, and probably going to repackage them into seperate units as there are 4 in each rackunit. The downside it the coloring these machines do though.

 

[midwayfair]

And you don't have to stick to 1M, you could use 5M or 10M, see what works best for your instruments.

What is the gain structure on the board you are going to? If the minimum gain is low enough you may not need to cut the level.

Do you know the output impedance on the keyboards you have? At least a buffer is probably warranted unless the keyboard has a really nice output stage with low impedance already. I would at least try the Schoeps circuit, it might just work fine for your keyboards.

That may not matter for you, the band I run sound for uses a few DI boxes, and the only one I need to lift ground on is a DI fed from the effects send of a bass amp, so that I can take the output of the bass preamp direct rather than try to mic the amp. I've never had to worry about ground lift with the Roland or Yamaha keyboards we have used, or the acoustic guitar pickups.
That is kind of a philosophical decision, do you want to make something which is as flexible as possible, or is just meeting your immediate needs good enough?

Roland, unfortunately, doesn't publish the output impedance of my (FP5) or my one of my friend's Rolands. Mine gets noisy if I run a long instrument cable but theirs doesn't. One of my friends plays a Nord .... uhh, one of the Stages (73? 74?), which I'm pretty sure I can run a thousand feet of cable from without it changing the sound.

I can also say that no instrument I've ever played live where we had access to direct boxes has been affected by a ground lift (meaning, if something was making noise, flipping a ground lift on the DI box on stage didn't change anything).

I'd like two of them to be as flexible as possible, whatever else I make can be an experiment.

 

[vintagelove]

Roland, unfortunately, doesn't publish the output impedance of my (FP5) or my one of my friend's Rolands. Mine gets noisy if I run a long instrument cable but theirs doesn't. One of my friends plays a Nord .... uhh, one of the Stages (73? 74?), which I'm pretty sure I can run a thousand feet of cable from without it changing the sound.

I can also say that no instrument I've ever played live where we had access to direct boxes has been affected by a ground lift (meaning, if something was making noise, flipping a ground lift on the DI box on stage didn't change anything).

I'd like two of them to be as flexible as possible, whatever else I make can be an experiment.

With regards to ground lift, I have had some issues with a keyboard plugged into something (an interface or an amp for instance), while simultaneously hooked up to a computer via usb. Lifting the ground helped.

Just something you might not want to skip if you build your own. Especially considering we’re likely to see more computers making their way into keyboard players workflows.

 

[Newmarket]

With regards to ground lift, I have had some issues with a keyboard plugged into something (an interface or an amp for instance), while simultaneously hooked up to a computer via usb. Lifting the ground helped.

Look at the parallel resistor capacitor solution to 'lifting ground'. Usually sorts it and allows phantom power. See Bo Hansen DI as an example.

 

[vintagelove]

Look at the parallel resistor capacitor solution to 'lifting ground'. Usually sorts it and allows phantom power. See No Hansen DI as an example.

I appreciate the tip. I had a cheap di laying around that I used to solve it.


Worth pointing out though, that Cheap DI… man that sucked tone so bad. Even for line level level, etc). Beware of those cheap units/transformers if building your own.

It was a RapcoHorizon SL-4. Avoid like the plague.

https://www.rapcohorizon.com/product/245/sl-4-rack-mount-di

 

[MicMaven]

t I could stick a FET follower in front.

Hmmm ... that sorta, kinda sounds inconsistent, ya know, putting a follower in front ... um ... er ...

(Nothing serious here, just poking fun at how the language works.) JR

 

[JohnRoberts]

Look at the parallel resistor capacitor solution to 'lifting ground'. Usually sorts it and allows phantom power. See Bo Hansen DI as an example.

I think this was called "hybrid" grounding scheme in the classic Ott textbook.

A low impedance at HF capacitor like a ceramic disc, looks like a short at RF frequency but an open circuit at mains frequency to prevent ground loops. A resistor provides a DC path. To support phantom powered microphones this R needs to be low single digit K ohms. For general grounding the R values can be higher.

I believe this technique was widely used in broadcast applications to keep RF noise in check.

JR

 

[Whoops]

Anyone have any suggestions? Projects, common transformers to use, or if this is a better-to-buy situation?

Yes, it's a Better to Buy situation.

if you want to spend more cash and waste a lot of your free time go for the DIY route and do the Bo Hansen DI

Thats it, done

 

[rivermAn]

Maybe the JLM FET DI kit would suit your needs. The entire kit is $44 AUD, easy to build, he's v helpful also:
https://www.jlmaudio.com/shop/fet-di.html?display_tax_prices=1