trinityamps.com

Hi everyone,

I received my OSD kit yesterday and figured it would be fun to document the build. I already have a couple of observations and questions!

First, the Trinity gang did a fantastic job putting this kit together. There was some damage to the box when it arrived and the power transformer was rattling around near the damaged section, but all of the contents were fine and unharmed - I can't blame Trinity for rough handling in transit!

Ok, on to my first issue!

Extra switch hole / missing switch
I've discovered is that there's a hole/label for a switch on the chassis (labeled "bypass" between the send and receive jacks) that isn't on either the schematic or the layout diagrams provided with the kit. Further, there are only three little red DPDT switches in the kit (mids, rock/jazz, 6v6/6l6) but I need another one for the Bypass switch. I guess I should reach out to Stephen for help?

Here's the panel:



And the inside view along with OSD Layout (revision 20) document:



And the section of the schematic (revision 14) that's switchless:



And here's the chassis (sans switch!) after the first build sprint:



Has anyone come across the switch issue on their builds?

Looks like my kit is missing the 500pF silver mica cap needed for V1. Luckily I have on in the bits box!

I am by no means an expert, but I believe that is there for a bypass switch if you install this: https://www.trinityamps.com/product/tube-effects-loop/

Ah, that makes perfect sense! Thank you I guess it's safe to leave it unpopulated or, if I want to get fancy, put the R/C filter on a switch.

When installing the 270pF ceramic caps on V2 I notice that the ceramic dielectric has crazy unstable dielectric when it comes to temperature fluctuations. This is the part I'm talking about:



At room temperature the ceramic cap measures within 20% of the stated value:



However if I warm the cap in my fingers to simulate the elevated temperatures inside a tube amp, the capacitance nearly doubles:



I have to assume that it would increase further with the hotter temperatures caused by glowing tubes! My question is this: has the vastly increased capacitance been figured into the design and is operating as desired/expected, or is the design based around a fairly stable 270pF, or does it not matter?

Edit: Looks like these are SL0 caps, which have a poor temp coefficient of around 120PPM. Something like a C0G/NP0 would get that down to 30PPM, much better. Same for mica.

I have some 220pF and 47pF silver mica caps that I could put in parallel for a much better temperature coefficient. Is this desirable/sensible?

Edit 2: Looks like those caps are for stability, preventing oscillations above the audio band. 220pF might be sufficient! I'll use those for now and check it out on the scope later. If more capacitance is needed I'll add them in parallel as required.

Thanks!

Here's a discrepancy from the BOM / layout / schematic: the 470R on pin 4 of both V4 and V5 is specced as a 5W part, but a 6.5W part is included in the kit. No biggie, I prefer the margin of error

Well, the new 470R are just big enough not to fit properly in the space between V4 socket and the 2-tab strip. Same for V5. As a result I had to bend the terminal strip to accommodate the larger component, then ran the resistor above the terminals instead of between them:



When it came to fitting the 220k // 500pF RC filter between Send and Return jacks I found that the supplied 500pF was a ceramic SL0 that measured 430pF. That's too far outside spec for my liking, so I used one of my own 500pF silver mica caps, which will also have better temperature coefficient and will therefore be more stable as a filter:

In the instructions at step 4 they say to connect jumper wires to the "ring" connectors. These are incorrectly labeled as "ring" and should in fact be "sleeve" or "shield", as they refer to the ground connection. The "ring" is only present on TRS (aka stereo) jacks.

Well, crap. I gone done f*cked up.

The instructions say to "Cut 6 - 2" long pieces of solid, 22 gauge, bare wire" so I took the ONLY piece of bare solid wire in the kit and cut it into 6x 2" pieces:



But it turns out that the only bare solid wire in the kit is actually the bus bar, or in my case, WAS the bus bar. I guess I didn't realize how thick "22 gauge" was

And now I have no bus bar and no solid, bare, 22 gauge wire. Crap.

EDIT: Yes! I found some old lengths of 12-2 mains cable in the garage, so I cut a length of the ground conductor out to use as my bus bar. It's thicker than the one included with the kit, but that's ok with me. Phew, saved the day! Please not people: check to see if your kit has the bare/solid 22AWG wire before you start to chop things up!!

So I'm really starting to worry about the ceramic caps provided in the kit. There's a 15pF that goes across pins 2&3 of the master volume pot, but despite the cap measuring 13pF at room temperature (perfectly fine), if I hold it in my fingers it quickly jumps up to 170pF!!! I fully believe the temperature of the chassis will exceed that of my fingers, so 200pF isn't an unreasonable expectation for where the master cap will end up.

Like the caps before it, I'm not going to fit the kit-supplied part and will instead use a silver mica from the bits box, this time a 22pF. It's a little higher than spec at room temperature, but under heavy heat will hold up much better.

When wiring up the output transformer to the 6v6/6l6 switch I see the instructions say to run a blue wire and a brown wire from the switch to V4 and V5, respectively. They also say to leave enough lead length to swap brown/blue if the speaker squeals at turn-on (I assume because of NFB polarity). Problem is, the off-cuts from the transformer primaries aren't long enough to reach V5 taking the route shown in the documentation.

There's blue wire I can use in bundle of wires provided by the kit, but there's no brown wire in there.

It's a good job I have solid core wire from a previous amp built, but what to other people do in this situation?

Looks like the 24k/2W carbon film resistor is missing from my kit and I don't have a spare in the bits box. How annoying!

After reading about another builder suffering an arcing problem with his 6v6/6l6 switch, I took particular care over this part of the build. First I wrapped the wire and soldered it carefully into position:



Then I applied just enough silicone dielectric grease to cover the entire solder joint and a little of the wire insulation:



Then I placed the heatshrink over the top:



I repeated this process for each connection to the switch, and then applied a thin coat of the silicone dielectric grease all around the heatshrinked connections to make sure everything was comprehensively insulated. Hopefully this will do the job!

Despite a couple of snafus along the way (for example inconsistencies in the OD trim pot wiring instructions/diagrams) progress is coming along nicely. I've deviated slightly from the BOM and made a couple of substitutions, such as using film caps instead of electrolytics (same capacitance, different dielectric) for V1 and V2 cathodes, and I'm using my own color scheme for the wiring. But overall I'm happy with progress!



The square red caps are Wima MKS2 10uF / 50V rated for 100°C and replace the axial 10uF electrolytics on V1's cathodes. I've used ~ 2.5uF of orange drops instead of 4.7uF on V2's cathodes; if I find the OD channel is bass light I'll add more capacitance, but I like my overdrive tones to have tight bottom end, so I'm going to suck it and see before making more changes. My hope is that the bass control will be able to compensate if needed.

Talking of component substitutions, I've omitted the 2 series chains of 3x 1N4007 diode rectifiers and instead made 2 series chains of 4x LQA06T300 ultra-fast, ultra-soft recovery diodes. Why?

Well, one of the best ways to cut down power supply noise is to avoid it entirely. One of the sources of noise in a guitar amp with solid state rectification is the reverse recovery spikes generated by diodes as they do their work. The spikes are very difficult to mop up with simple R/C/R/C style filters, so I wanted to use something that is as quiet as possible out of the gate.

To that end, the LQA series of diodes (https://www.mouser.com/datasheet/2/328/ ... 512439.pdf) have an exceptional 9.3ns reverse recovery time and 19nC reverse recovery charge. I've used them to great effect in hi-fi audio amplifiers and fully expect them to work well in this application, too.



The model I've used is rated for only 300V peak repetitive inverse voltage, which is why I've used chains of 4 for 1200V PIV. I'm comfortable with that. If it blows up, I'll come back here to post pictures of the shattered and burned remnants of my amp.

More progress



I wired the 5V/GND wires from the 5-pin DIN connector slightly differently to the instructions, which state that the ground wire should be connected to the PRE1 point on the chassis. Instead I kept the 5V and GND wires tightly twisted and connected them both to the relay board right next to the smoothing capacitors in order to reduce loop area. See red/black wires in this photo:



I've also diverged slightly from the ground wires linking the power supply caps. This is one of the most critical areas of the amp for ensuring low impedances, and to that end I used 14AWG solid copper wire right against the capacitors GND terminals, like so:



I've tested the main board's power arrangements using a bench supply at low voltage (just 5V) and everything is measuring perfectly. Not long before I get to switch it on!

And she's done! Powered up great first time, bias for the 6L6s is perfect at 32mA and there appear to be no problems. The controls all seem to be working just right, foot switches are doing foot switchy things, OD channel controls working fine, presence is subtle but works great. Voltages are on the low side (B+ is ~ 410V), but this is to be expected with the LQA06T300 diodes, which have a much larger forward drop than the 1N4007s.

I've played a bit through my junk test speaker and there's barely perceptible hiss/hum (unless its cranked, of course) when idling. It sounds fantastically spanky clean! I still need to adjust the PI voltage offset and then I'll get it hooked up into a proper cab and have some fun.

But for now... job done!! Thanks very much to the Trinity team for a great kit and prompt support.

Here's she is in all the glory!

After some playing I found the OD channel a bit spitty and unpleasant, with microphonics that sounded just bad. So I swapped out the OD tube (a NOS Tesla / Westinghouse E83CC clone) for a NOS Westinghouse 12AX7A and it fixed the microphonics, but not the unpleasant spittiness of the OD.

Given that I'd messed with the cathode caps (using film 2.2uF instead of electrolytic 4.7uF supplied in the kit) I replaced them with some wonderful old axial wet tantalum caps I've had in the parts drawer for special occasions. They're 10uF/63V, so I wired two in series for each of the cathodes and they came out to 4.735uF each. Perfect!

OD now sounds spectacular. Caps are hooked up as follows:



You can also see a white rectangular cap in the photo above. It's a 1uF/50V Evox Rifa SMR (polyphenylene sulphide dielectric) that I'm using instead of the tant in the presence control.

Edit: I've removed bad advice.

So far so good! Very much enjoying the OSD

The cabinet arrived from Stephen and co. a couple of days ago and I have to say that it looks really, really good. I took a couple of pictures.











Full resolution photos here: https://imgur.com/a/2N1RteG

The only niggle I have is that the bolts attaching the amp to the cab are silver and are clearly visible on the top of the amp (also they don't come with washers, so I ended up screwing up against the tolex, which isn't ideal). All the other bolts/screws are black and it messes with the aesthetic, so I ordered some black #10-32 bolts and washers from Amazon. Job done!

You'll notice there's no knob on the impedance selector. This is deliberate and is intended to make it virtually impossible to accidentally change the impedance from 8 Ohms to something else; the rotating switch is extremely difficult to turn without a knob attached, or without pliers.

The purple knobs are anodized aluminum from Small Bear, and the purple jewel cover is from Mojotone. The speaker is an 8 Ohm, 60-Watt Fane Ascension A60, a super clean and very balanced alnico driver perfect for the enormous headroom in the Trinity OSD. Speaker cable is Mogami W3082 coax and the connector is a Redco right-angle TS.

Bolts: https://www.amazon.com/gp/product/B00FTJZ8Q8
Washers: https://www.amazon.com/gp/product/B009OJHQMI
Knobs: https://smallbear-electronics.mybigcomm ... cal-knurl/
Jewel: https://www.mojotone.com/amp-parts/Pilo ... rple-Jewel
Speaker: https://www.fane-acoustics.com/view-pro ... ension-A60
Cable: http://www.mogamicable.com/category/bul ... ure_sound/
Connector: https://www.redco.com/Redco-Right-Angle-Flat-TS.html