Tda1545a vref capacitor

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
ThorstenL

ThorstenL

Previously: Kuei Yang Wang
Joined 2002
linkedin.com/in/thorsten-loesch-studio-raumklangg/
Konnichiwa,

ash_dac said:
For the vref capacitor I have some rubycon mbz low esr 1000uf caps lying around so I modelled the circuit in SwitcherCad, and it "modelled" ok so may try this cap out today or tomorrow.
Click to expand...

Question is, will it fit the PCB? Otherwise sounds fine.

ash_dac said:
Thanks for the pointers towards upgrading the power supply. I thought I misread the diagrams at first (+10/-10V supply for opamps was unregulated) but nope it's all there!

Oh dear:dead:
Click to expand...

Yes, this has been the same since the CD-720/CD-36 (Philips/Marantz respectively) through all the versions.

I am not sure if the 723 now has a real crystal, the 720 used a (slightly) cheaper ceramic resonator.

My upgrades where LM6182 in the analogue stage (with the required upgrades in the PSU) new clock, seperate powersupplies (4pcs) for clock, dac and +/- Analog stage, -rm-rf the headphone section.

I eventually tried a 500VA external torroid transformer for the analogue powersupply it did give another little upgrade ove the internal 25VA EI Transformers I used origianlly.

That modded 720 was a tough customer!!!!

Sayonara
 
J

jzono1

Member
Joined 2005
Getting started - getting in

If you have not already got one, I strongly recommend that you get hold of a service manual for the CD723 from:
Willow Vale Electronics Limited
Connect Business Park
Bordesley Green Road
Birmingham
B9 4UA
England
Tel: 0870-6000-271
E-mail Willow Vale

Or click HERE for their web site.

Apparently, Willow Vale have not been contactable recently so I am indebted to Mike Mears for locating another source, namely Dutch West Distributions, tel - 01434 32 32 32. The manual appears to be slightly cheaper from them.
Click to expand...
-- http://myweb.tiscali.co.uk/nuukspot/decdun/project.cd2.html#cdstarting2

So, from where can I get a service manual then? Not good enoug in English to call 'em :-/
 
A

ash_dac

Member
Joined 2005
Kuei Yang Wang said:
Konnichiwa,



Well, the full scale current of the TDA1545 AND the bias (offset) current are both directly proportinal to Vref, with around 10db gain IIRC, so in effect -10db PSRR on Vref.

The PSRR on the supply line of the TDA1545 is listed as 30db.

In other words, unless both Vref and Vdd are squeaky clean, Vref to 18Bit levels and Vdd to 12bit levels the TDA1545, referenced to the output the chips performance will degrade, potentially very seriously.

Sayonara
Click to expand...


2) Completely galvanically seperate supply for the DAC, given that the DAC has crappy PSRR on both Vref and Vdd pin you cannot do "too much" here.
Click to expand...

Hi,

So that's the Vdd dealt with which is a large chunk of that equation from the tda1545a datasheet now to the resistors!

For the reference resistors (R3 and R4) I assume I choose precision 0.1% resistors with 'good' temperature coefficients ?


Some calculations:- (warning these may need correcting)

Ibias = Iref X Abias

So from the datasheet 714ua / 9.42 = 75.80ua needed for Iref.



From the resistors used in the datasheet:-

11K / (33K+22K+11K) x 5v = gives 0.83V across internal 11K Rref

(0.83/11000) X 1000000=75.46ua Iref



Trying 14.6K, 7.4K....

14.6K / (33K + 22K + 14.6K) X 5 =gives 1.04V across internal 14.6K Rref

(1.04/14600 )x 1000000= 71.23ua Iref


7.4K /(33K +22K +7.4K) X 5 = gives 0.59V across internal 7.4K

(0.59/7400) X 1000000 = 79.73ua Iref


Well done Philips give people like me practice in building power supplies :)

The datasheet says ,"Easy application: single 3 to 5.5 V rail power supply and output- and bias current are proportional to the supply voltage"

Says it all really!


I guess setting vref at 1/2 vdd with passive i/v (see link) you get more current output across the output resistors pushing the output voltage out of compliance causing distortion.


http://diyparadise.com/dackit/1545dackit.html

Please correct me on my calc's if i am wrong.



Kind regards,


Ashley.
 
hollowman

hollowman

Banned
Joined 2007
There is probab. a simple answer for this question but I'm too lazy to search ;) ...
For Vref (Iref), Philips, in the datasheet schematic, suggests a 2/3 Vdd voltage divided be created using 22k and 33k resistors.
Given the small voltage (<5vdc) and currents in this topology, why are THESE relatively "high"-value resistors suggested (as opposed to lower values)?
E.g., is it better damping?
Etc...
An externally hosted image should be here but it was not working when we last tested it.
 
Last edited:
abraxalito

abraxalito

Member
Joined 2007
The resistor values are high because the internal resistor of the 1545 is also relatively high (11k nominal). So the divider resistors are chosen to give 1/6th of the supply voltage across the internal resistor, at the nominal value. However it has a fairly wide tolerance.

Its possible to use lower value resistors but then you'd need to put a resistor in parallel with the internal one (pin7 to pin4) to shunt a proportion of the current away. Can't see any value in that though - you'd also need a larger value of decoupling cap (C2 in your schematic) to achieve similar PSRR.

<edit> I just realized this was assuming 1mA output current - the DAC will accept up to 2mA, when there will be 1/3rd of the supply across the internal 11k, corresponding to 1.66V nominally.
 
Last edited:
hollowman

hollowman

Banned
Joined 2007
I wasn't too sure that "internal" R_ref=11k -- shown in Fig. 1 -- was real or a parallel equivalence.
That is, in the DS (pg. 9, Note #5), it states:
R3=R4=11k
...and this supposedly gives the highest S/N in the parametric table above that Note.

I have no idea what Note 5 means with that remark R3=R4=11k.

Also, a bit weird (as are a LOT things about this DS and 1545 device!!) is putting a fixed-value R (R_ref, 11k, if that's a real resistor) *inside* the device where a developer has no control (esp. WRT to quality, e.g., tolerance, metal film, etc.)

EDIT: I measured a naked TDA1545a (it is a "true"/genuine/OEM Philips chip, AFAICT, w/o "TAIWAN" label) at Pin 7 WRT Ground (pin 4) using a Fluke 87 DMM:

10.15k-ohm
 
Last edited:
hollowman

hollowman

Banned
Joined 2007
Another thing ...
If one is using NON-DIFFERENTIAL inputs on the OP, as in the case of popular Rudolf I/V stage, then there no (+) OP1 and OP2 input for V_ref to connect to (see DS, Fig. 1). V_ref, hence, "sees" infinite resistance/impedance.
In this case, I wonder if it's not helpful to have SOMETHING for V_ref to sink to -- i.e., a very-high-value R, or even a snubber ?
 
hollowman

hollowman

Banned
Joined 2007
R_ref and datasheet V_ref (I_ref) R3, R4

hollowman said:
I measured a naked TDA1545a (it is a "true"/genuine/OEM Philips chip, AFAICT, w/o "TAIWAN" label) at Pin 7 WRT Ground (pin 4) using a Fluke 87 DMM:
10.15k-ohm
Click to expand...
...and with unpowered 1545 per the datasheet ckt (Fig. 1), using
R3=22k, R4=33k ...
V_dd to GND=64k
I_ref (pin 7) to GND=9.5k
V_dd to I_ref (pin 7)=55k (or: 22k + 33k, as expected)
V_ref to GND=41.5k
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom