DT50 input clipping - anybody ideas?

[Angelo]

HI,

 

After years of working great, my DT-50 has extremely harsh clipping on the guitar input.  

 

Symptoms:

Plugging my guitar straight into the amp, I get clipping when playing chords at a normal volume.   If I turn my guitar volume way down I don't experience any clipping at all.   It seems like a bad diodes or bad op-amp.

 

Things I tried:

I connected the DT-50 via L6 Link with my HD Pro rack; when I do this the DT-50 sounds perfect.   It is definitely something related to the guitar input.

Changed the tubes on the DT-50 - no difference at all.  It wasn't the tubes.

 

Input Board part number:

Opening up the amp, the guitar input board is over on it's own.   The part number on the input board is 35-00-0385-1     The input board has the TLC084C opamp package on it.    You can but the DT-25 input board on Full Compass, but the DT-50 input board is not for sale on Full Compass.     I'm suspecring the 

 

Any ideas,  Anybody else experience the same thing?

 

[psarkissian]

How long since the tube were replaced?

Op-amp? Maybe. But if it sounds fine thru L6 Link, then the op-amp is good. So I suspect one of the pre-amp tubes

is getting old, the power tube are getting old and negative feedback is not moderating the gain, or the all the tubes are getting old.

 

[Angelo]

Hi Partev!

 

I just replaced all four tubes in the DT-50

 

Using L6 Link means that the guitar input board is not used, and the digital audio goes into the DSP and then to the tube power amp.   

 

Since the DT-50 functions perfectly when L6 Link is being used, it means that the DSP and the power amp are not where the problem is.

 

If you open up the DT-50 you will see that there is a daughterboard for the guitar input that is separate from the mainboard.   The part number for the daughterboard is 35-00-0385-1

 

Do you know where I can get a replacement for the daughterboard to try out?  

 

 

[psarkissian]

BTW, the part number is wrong, that's the un-stuffed board number. I don't see them on

Full Compass, so I don't think they are still available.

 

Input board goes to the main board, that goes to the L6 Link interface, the model signal

bypassing the tube stage with the modeled signal.

 

What tubes were put in, and what bias???

 

 

[Angelo]

A pair of Ruby 12AX7AC5 HG+ tubes and a pair of Sovtek tubes for the power amp.

Changing the the tubes made no difference at all in the performance, so I swapped back.

 

I don't think I am describing my test cases clearly enough.

 

My L6 Link signal flow - no input clipping - sounds great:

Gibson SG -> HD Pro rack -> L6 Link -> DT-50     

In this case the input board of the DT-50 is not being used

 

My DT-50 signal flow - input clipping - sounds bad:

Gibson SG -> DT-50

In this case the input board of the DT-50 is being used

 

It is only when I connect my guitar to the input of the DT-50 that I get the weird lack of headroom and clipping

 

[Angelo]

Also, what is the correct part number?  If 35-00-0385-1 is the unstuffed pcb, then what is the fully populated pcba?   My guess is 55-02-0385, but just a stab in the dark based on the DT-25 input board number.   

 

Thanks for all the speedy replies 

[psarkissian]

50-02-0385-1, but I couldn't find any out there, I looked.

High Gain Ruby,... no, don't use that. These are not all analog tube amp of the 1960's and 70's, so can't go swapping

like we used to back in the day. Follow the choices in the link, our tube amps are finicky that way. Goes with being a

hybrid analog/digital tube amp.

 

 

Make certain it's correctly biased, take it to an authorized Line 6 service for that, there is a bias and burn-in

procedure to be done on those after tube replacement.

 

[Angelo]

Ok, thanks for the part number and the tube tips.

 

Aren't the 12AX7's after the DSP main board?  I am nearly certain that they are.

 

 

 

[psarkissian]

Yes, program board out to a circuit, then the pre-amp tubes.

[Angelo]

Ok great, then that exonerates the tubes.   The tubes are working perfectly.

 

If I go:  Gibson SG -> HD Pro rack -> L6 Link -> DT-50, I have no issue at all

 

There is clipping somewhere between the DT-50's guitar in and the codec of the DSP on the mainboard.

[psarkissian]

Using the correct tubes and bias?

 

 

[Angelo]

yes, correct tubes and bias

 

I have no clipping if the HD Rack does the conversion of the guitar signal from analog to digital, and then sends the digital audio to the DT-50 via L6 Link.

 

I only get the clipping if I plug my guitar directly into the front of the DT-50.

[psarkissian]

High or Low input?

 

[Angelo]

Both

[psarkissian]

Hmmm,... yeah, that's sounding like an amp chip in the input alright.

 

 

[Angelo]

yeah, that's what I was thinking of trying first.  The part # is TLC084C.    I was posting to see if it was a common problem, but it looks like it might only be me.  :) 

[psarkissian]

Rarely have a problem with that. TL08x series has differential JFET inputs that are high enough impedance so that transients don't

damage the chip. So they are used for in's and outs on audio devices, as opposed to a 5532/34 which are bipolar differential inputs,

and more susceptible to transients. But since they are lower input noise spec, are used for internal audio signal circuits in a host of

devices like the old SSL 3000 and 4000 analog series of days past (still a favorite among many audio engineers, including me). And

I design audio circuits using a lot of the 5532/34 series and LM833 (Nat'l Semiconductor equal), as input noise and slew rate are the

two of the more important specs for my uses.

 

A lot of Celine Dion's "Brand New Day" and Glenn Hughes recordings were done on a 4000 series at Westlake Audio Recording,

Studio-C,  back in the day (2001-2002).

 

[Angelo]

Interesting.

 

If it turns out that the TLC084C is okay.  Are there other parts in the input audio path that could contribute to severely lower headroom if they fail?    Diodes maybe?

[psarkissian]

Not that I recall. So rare that it goes out like that.

 

[Angelo]

I haven't been able to swap out the TLC084C to narrow down the issue yet.

 

Is the DT-25 guitar input board compatible with the DT-50 (electrically)?   They have the same header

 

https://www.fullcompass.com/prod/270207-line-6-50-02-0209-guitiar-input-pcb-assembly-for-dt25-amp

[psarkissian]

Do not mix and match between the DT50 and DT25. 

 

[Angelo]

Okay, I've swapped out the TLC084C.   That did not fix the issue, but I have more information.

 

Coming out of the DT-50's input board are differential signals for a High Gain and Low Gain signals that go to the codec.   These differential channels go to the codec are used for gain switching in the DSP.   When you are playing quietly the DSP uses a peak follower to select the High Gain differential channel to keep your guitar input signal level higher than the noise floor of the electronics.   When your input crosses some threshold, the DPS selects the Low Gain differential channel to provide you with more headroom.   This gives you the best of both worlds, low noise when you need it, and higher headroom when you need it.

 

This tracks exactly with my problem.   When I play quietly, the guitar sounds crystal clear (I'm on the High Gain differential channel), but when I dig in to the strings the signal cuts out and clips, which means there is a problem with the Low Gain differential channel, or possibly an issue with the Codec (a Cirrus Logic CS42436).   I highly doubt just a single channel went bad on the codec, I've never heard of that happening.   

 

Partev, do you have any clues?   Have you had any other DT-50 failures with the input gain switching?     I'll keep probing to see if I can find the fault.   I hate to have to scrap my DT-50 when it is otherwise working perfectly!

 

Best,

Angelo

[psarkissian]

"Have you had any other DT-50 failures with the input gain switching? " --- Gain switching on a CoDec,...

a bit more involved than that. There is a standard set of circuits used on the front end of CoDec's.

 

So Angelo,

Op-amps are set-up in +In, -In to balance inputs to a CoDec, aka- reference drivers or line drivers. Typical

with chip company's application notes. Cirrus, who also acquired Cypress some years back, and companies 

like AKM will have these kinds of front en circuits in their app notes. As well as a similar circuit on the

analog out of a Codec.

 

The old ADAT's used Cypress Codec IC's back in the day, before Cirrus acquired them. High quality

20-bit back then when I service tech'd for Alesis 20-years ago (geez, has is been that long?).

 

And Angelo,... also check out a paper on current steering DAC outputs for high quality audio (by my pal) Malcolm

Hawksford at Audio Research Lab, University of Essex, UK. Love his work.

 

https://www.researchgate.net/publication/269102687_CURRENT-STEERING_TRANSIMPEDANCE_AMPLIFIERS_FOR_HIGH-RESOLUTION_DIGITAL-TO-ANALOGUE_CONVERTERS

 

When using the correct of the two inputs input gain and clipping are a problem, and it's not related to

the guitar being turn up full-on, the main reasons tend to be the input board, the tubes and/or tube bias.

 

[Angelo]

Still trying to track the issue down.

 

Just to summarize.   This is not an issue with the input board, this is not an issue with the tubes, and this is not an issue with the tube bias.

 

There are two inputs on the front of the amp, and they are both working as normal to the input board.    The part where the nomenclature gets confusing are the signals coming out of the input board.    I'm looking at the signals GIN_LO+, GIN_LO-, GIN_HI+, and GIN_HI-      There is a 20x gain difference between the two signals.   If my memory serves me correctly, this is used for the ADC Crossfade Gain Switching.    The way that works is the DSP is listening to both signals and uses the high gain channel at low levels to provide better SNR and then switches to the low gain channel at higher input levels to provide better headroom.    What I believe is happening is that the DSP is never switching over to the GIN_LO+, GIN_LO- pair so I never get the benefit of the higher headroom.

 

In the pictures, I am probing the test points between GIN_LO+ and GIN_HI+, for a signal of 290mV RMS on GIN_LO+, I get 4.7V RMS on GIN_HI+.   This is right at the edge of clipping for the GIN_HI+ channel.   To me, this exonerates the input section up to the Codec.   So now I suspect the Codec may have issues, but I have never experienced just a single channel going bad on a Codec.

 

Again, I hate to scrap my DT-50.   It works perfectly when used through Line 6 link with my POD HD.    It is just that the input switching isn't working properly

 

Best,

Angelo

 

[psarkissian]

-Could be amp at input or output of CoDec.

-Chip on the amp board.

-Could be master volume is pulled up and in "pad" mode.

 

Does it motor boat when you use the master volume control?

 

[lanceaustin]

Angelo, did you ever figure this out?

 

My DT50 appears to be doing the same thing. 

[psarkissian]

About four different things it could be. Have troubleshooted by a Line 6 authorized service center in your area.

 

https://line6.com/find/service_center/

 

 

[FFRUST]

On 1/3/2025 at 7:26 PM, lanceaustin said:

Angelo, did you ever figure this out?

 

My DT50 appears to be doing the same thing. 

 

Hi Ianceaustin,

 

Did you get yours fixed? Mine is doing the same.

 

 

[FFRUST]

@Angelo @lanceaustin

 

Any of you got your DT50's sorted?

 

I have the same issue.

[astrahle]

I am also experiencing the exact same issue with my DT50 112. Works perfect through L6 Link at any volume, irrespective of 12AX7 boost on/off, or B+ Voltage low/high. When connected to either High or Low input on the DT50, quiet playing is fine, however get harsh digital clipping when providing high transients. Have replaced all 4 tubes with Electro-Harmonix tubes however problem persists

[psarkissian]

Are you using active pick-ups? 

 

[astrahle]

Ok, getting all 2025 on this problem's arse>

Here's what AI has to say after referencing this post and the DT50 schematics. I'm handy with a soldering iron, but prefer to rely on the nouse of an electrical engineer/amp boffin. If anyone can help and advise any progress, would be a great help to all.

This is a classic “front-end only” failure on the DT50, not a tube/power section problem.

 

Because L6 LINK audio works perfectly (and still runs through the 12AX7 PI and EL34s), the fault has to be before the A/D stage that feeds the DSP—i.e., in the analog input path to the codec (or the codec itself), not in any of the tube stages.

 

 

What that narrows it to

 

 

On the DT50 the two front-panel inputs (Hi/Lo) go through:

 

1. jack/pad & ESD/protection →

2. a small analog buffer/gain stage →

3. coupling into the A/D of the audio codec on the digital/main board →

4. DSP → DAC → tube phase inverter → EL34s.

 

 

L6 LINK injects digital program material after step (3). So if L6 LINK is clean but both front inputs harsh-clip at modest level, the likely culprits are:

 

• Leaky/shorted input clamp/TVS diodes on the main board side (not just at the jack board). When one of the back-to-back clamps goes partially short, you get squarey, “diode-like” clipping at very low level.

• Failed/browning analog rail for the input op-amp(s) (±12/±15 V or +5 V analog, depending on the exact rev). If an op-amp is running on one rail or a sagging rail, it clips way early.

• Lost mid-bias (virtual half-supply) to the codec input. If the codec expects ~½ AVDD bias and that node has drifted (bad resistor or dried cap), the headroom collapses and you get asymmetric, harsh clipping.

• Bad coupling electrolytic feeding the codec (ESR gone high or DC leakage), causing DC offset and immediate overload at normal guitar levels.

• The audio codec’s A/D input pin itself (damaged from ESD/lightning/static through the guitar lead). A partially blown input on an AKM/Cirrus codec presents exactly like this: normal at whisper levels, then ugly crunch once you hit a few hundred mVpp.

• Much less likely (but checkable): a software/global pad setting or corrupted calibration—but your description (hard, immediate clipping) screams hardware, not gain mis-set.

 

 

 

Why it’s almost certainly not the jack “input board”

 

 

Someone on the Line 6 forum replaced that board and still had the problem. That board mainly carries the jacks, pad resistors, and initial ESD parts—so the fault is very probably downstream on the main/DSP board (or its power rails).  

 

Also, the DT50 documentation confirms L6 LINK carries audio digitally, bypassing the analog guitar input A/D path—matching your observation that Link is clean at any level.  

 

 

Focused diagnostics (safe, quick checks first)

 

 

⚠️ Safety: this amp has lethal voltages even after power-off. If you’re not fully comfortable with tube-amp safety, hand it to a tech. Bleed caps, one-hand rule, isolation, etc.

 

1. Measure the low-voltage rails on the main board (no guitar plugged):

    

   • Check the analog op-amp rails (often ±12 V or ±15 V) and the codec rails (AVDD 5 V or 3.3 V, DVDD 3.3 V). Anything more than ~5% off, or big ripple, = suspect regulator or filter caps. Early clipping with both inputs points to this.

    

2. Scope the signal path with a 100–200 mVrms 1 kHz test tone (or your guitar very lightly picked).

    

   • At the last op-amp output before the codec input: If it’s already flattening, the op-amp/rail/bias is bad.

   • If it looks clean there but flattens right at the codec input, suspect the clamp diodes or the codec inputitself.

    

3. Mid-bias node check: Most codecs expect ~½ AVDD bias at their analog inputs. Find that reference network and verify it’s at the correct voltage and stable. A drifting bias cap will make the waveform slam into one rail.

4. DC at the codec input: With no signal, you should see the expected mid-bias DC only. Any extra DC (tens to hundreds of mV) hints at a leaky coupling cap or a damaged input stage.

5. Lift/replace the clamp network: If there are back-to-back diodes or an IC-style ESD protector right before the codec, temporarily lifting one leg (for diagnosis only) will tell you fast if a leaky clamp is compressing/clipping the input. Replace as needed.

6. Reflow the board connectors between the jack board and main board. Micro-cracks can upset bias/ground reference and cause early clipping.

7. If all pass, replace the audio codec IC (and any nearby electrolytics). This has fixed identical “low-level harsh clipping” faults on other Line 6 units over the years.

 

 

 

Likely parts to earmark (by function)

 

 

• Protection/ESD diodes (SOT-23 or SOD-123 near the audio in header): replace as a set.

• Coupling electrolytics in the few µF–tens of µF range feeding the codec input(s).

• Analog op-amp in the input buffer (often a TL072/TL074-class part or low-noise single/dual); replace + inspect its rail decouplers (100 nF ceramics + 10–47 µF electrolytics).

• Low-voltage regulators for ± rails and 3.3/5 V analog (and their filter caps).

• The codec itself (AKM/Cirrus part adjacent to the DSP).

 

 

 

Why these are the prime suspects

 

 

• Your symptom is hard, non-musical clipping at moderate pick attack, on both front inputs, while Helix via L6 LINK is fine. That’s textbook “front-end analog headroom is gone” or “A/D input clamped”, not tubes or power amp.

• Replacing only the jack/input PCB didn’t help in another identical case → points to main board analog/A-D path, not the jacks.  

• The DT50’s L6 LINK design explicitly bypasses the analog guitar input and A/D. If the clipping were anywhere after the codec (DSP, DAC, tube path), L6 LINK would clip too—but it doesn’t.  

 

 

 

If you want a fastest practical repair path

 

 

1. Replace the coupling caps into the codec and the input clamp/ESD diodes (cheap, common failure points).

2. If unchanged, replace the input buffer op-amp (and confirm its rails/decouplers).

3. If still unchanged, swap the codec IC. (Yes, it’s SMD; any competent audio tech can do it.)

4. Along the way, verify/recap the small analog rail filters for the input section.

 

 

If you want, tell me the reference designators around the guitar-input header on the main board (from your schematic: the codec “Uxx”, the nearby diodes “Dxx”, the buffer op-amp “Uxx”, and the coupling caps “Cxx”), and I’ll map exact test points and voltages to check.

 

Useful references: Line 6 forum thread with same symptoms (input board swap didn’t help), and DT50 docs explaining how L6 LINK carries audio digitally beyond the analog input stage.  

 

Good luck—and please be careful around the HV! If you can share a couple of scope screenshots (op-amp out vs codec in at the onset of clipping), I can pinpoint which node is failing.